The Balance Evaluation Systems Test

The Balance Evaluation Systems Test(BESTest) to DifferentiateBalance DeficitsFay B Horak, Diane M Wrisley, James Frank

Background. Current clinical balance assessment tools do not aim to help ther-apists identify the underlying postural control systems responsible for poor functionalbalance. By identifying the disordered systems underlying balance control, therapistscan direct specific types of intervention for different types of balance problems.

Objective. The goal of this study was to develop a clinical balance assessment toolthat aims to target 6 different balance control systems so that specific rehabilitationapproaches can be designed for different balance deficits. This article presents thetheoretical framework, interrater reliability, and preliminary concurrent validity forthis new instrument, the Balance Evaluation Systems Test (BESTest).

Design. The BESTest consists of 36 items, grouped into 6 systems: “BiomechanicalConstraints,” “Stability Limits/Verticality,” “Anticipatory Postural Adjustments,” “Pos-tural Responses,” “Sensory Orientation,” and “Stability in Gait.”

Methods. In 2 interrater trials, 22 subjects with and without balance disorders,ranging in age from 50 to 88 years, were rated concurrently on the BESTest by 19therapists, students, and balance researchers. Concurrent validity was measured bycorrelation between the BESTest and balance confidence, as assessed with theActivities-specific Balance Confidence (ABC) Scale.

Results. Consistent with our theoretical framework, subjects with different diag-noses scored poorly on different sections of the BESTest. The intraclass correlationcoefficient (ICC) for interrater reliability for the test as a whole was .91, with the 6section ICCs ranging from .79 to .96. The Kendall coefficient of concordance amongraters ranged from .46 to 1.00 for the 36 individual items. Concurrent validity of thecorrelation between the BESTest and the ABC Scale was r�.636, P�.01.

Limitations. Further testing is needed to determine whether: (1) the sections ofthe BESTest actually detect independent balance deficits, (2) other systems importantfor balance control should be added, and (3) a shorter version of the test is possibleby eliminating redundant or insensitive items.

Conclusions. The BESTest is easy to learn to administer, with excellent reliabilityand very good validity. It is unique in allowing clinicians to determine the type ofbalance problems to direct specific treatments for their patients. By organizingclinical balance test items already in use, combined with new items not currentlyavailable, the BESTest is the most comprehensive clinical balance tool available andwarrants further development.

FB Horak, PT, PhD, is ResearchProfessor of Neurology and Ad-junct Professor of Physiology andBiomedical Engineering, Depart-ment of Neurology, OregonHealth and Sciences University,West Campus, Building 1, 505NW 185th Ave, Beaverton, OR97006-3499 (USA). Address allcorrespondence to Dr Horak at:[email protected].

DM Wrisley, PT, PhD, NCS, is As-sistant Professor, Department ofRehabilitation Science, Universityat Buffalo, Buffalo, New York.

J Frank, PhD, is Dean of GraduateStudies, Department of Kinesiol-ogy, University of Windsor, Wind-sor, Ontario, Canada.

[Horak FB, Wrisley DM, Frank J.The Balance Evaluation SystemsTest (BESTest) to differentiate bal-ance deficits. Phys Ther. 2009;89:484–498.]

© 2009 American Physical TherapyAssociation

Research Report

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

484 f Physical Therapy Volume 89 Number 5 May 2009

Balance deficits are one of themost common problems treatedby physical therapists. Thera-

pists need to identify who has a bal-ance problem and then decide thebest approach to rehabilitation. Cur-rent standardized clinical balanceassessment tools are directed atscreening for balance problems andpredicting fall risk, particularly in el-derly people.1–7 These tools identifywhich patients may benefit from bal-ance retraining, but they do not helptherapists decide how to treat theunderlying balance problems. Be-sides not being aimed at guidingtreatment, the current balance as-sessment tools were developed spe-cifically for older adults with balanceproblems. This article presents anew balance assessment tool devel-oped to help physical therapistsidentify the underlying postural con-trol systems that may be responsiblefor poor functional balance so thattreatments can be directed specifi-cally at the abnormal underlyingsystems.

Although many clinical tests are de-signed to test a single “balance sys-tem,” balance control is very com-plex and involves many differentunderlying systems.8–11 Whereasprevious motor control models as-sumed postural control consisted ofheirarchical righting and equilibriumreflexes, we wanted to develop aclinical test of balance control basedon Bernstein’s concept that posturalcontrol results from a set of interact-ing systems.11–16 Consistent with this“systems model of motor control,”recent research in our laboratory andothers has demonstrated how con-straints, or deficits, in different un-derlying systems can impairbalance.10,11,13,15,17–20

Constraints on the biomechanicalsystem, such as ankle or hip weak-ness and flexed postural alignment,limit the ability of frail elderly peopleand patients with Parkinson disease

(PD) to use an ankle strategy or com-pensatory steps for postural recov-ery.21,22 Constraints on the limits ofstability (that is, how far the body’scenter of mass can be moved overits base of support) and on vertical-ity (that is, representation of gravita-tional upright), affected by sensorydeficits or by stroke in the parietalcortex, may result in inflexible pos-tural alignment or precarious bodytilt.23,24

Constraints on anticipatory pos-tural adjustments prior to voluntarymovements depend on interaction ofsupplementary motor areas with thebasal ganglia and brain-stem areasand result in instability during stepinitiation or during rapid arm move-ments while standing.25,26 Con-straints on short, medium, and longproprioceptive feedback loops re-sponsible for automatic posturalresponses to slips, trips, and pushesinclude late responses in patientswith sensory neuropathy or multi-ple sclerosis, weak responses in pa-tients with PD, and hypermetric re-sponses in patients with cerebellarataxia.27–31

Constraints on sensory integrationfor spatial orientation result in disori-entation and instability in patientswith deficits in pathways involvingthe vestibular system and sensory in-tegrative areas of the temporoparietalcortex when the support surface orvisual environments are moving.27,32,33

Constraints on dynamic balanceduring gait result from impaired co-ordination between spinal locomo-tor and brain-stem postural sensori-motor programs when the fallingbody’s center of mass must becaught by a changing base of footsupport.34 In addition, cognitive con-straints on executive or attentionalsystems can compound constraintsin the other systems because eachunderlying neural control system forbalance control requires corticalattention.12

Figure 1 shows the 6 interacting sys-tems underlying control of balancethat are targeted in our new BalanceEvaluation Systems Test (BESTest).Each system consists of the neuro-physiological mechanisms that con-trol a particular aspect of posturalcontrol. Many of these systems areindependent from each other in thatdifferent neural circuitry is involved,such that different pathologies mayinvolve damage to different systems.For example, people with PD mayhave an abnormal system for step-ping in response to an external per-turbations but a normal sensory ori-entation system, which allows themto stand with eyes closed on an un-stable surface by relying upon vestib-ular information.27,35 In contrast,people with loss of peripheral ves-tibular inputs may have abnormalsensory orientation with eyes closedon an unstable surface but normalpostural responses to external per-turbations.36,37 In current practice,computerized, dynamic posturogra-phy is based on the concept thatthe sensory orientation and posturalmotor reactions systems underlyingbalance can be separately measuredand represent separate systems un-derlying control of balance.38 Thus,each patient with balance problemsis likely to fall because of deficits indifferent underlying systems andmay consequently fall in differentenvironments and while performingdifferent tasks. Therapists need tobe able to differentiate the under-lying systems’ contribution to bal-

Available WithThis Article atwww.ptjournal.org

• eAppendix: Balance EvaluationSystems Test (BESTest)

• Audio Abstracts Podcast

This article was published ahead ofprint on March 27, 2009, atwww.ptjournal.org.

Balance Evaluation Systems Test (BESTest)

May 2009 Volume 89 Number 5 Physical Therapy f 485

http://www.ptjournal.org/cgi/content/full/89/5/484/DC1

http://www.ptjournal.org/cgi/content/full/89/5/DC1

ance problems and fall risk in theirpatients in order to appropriately di-rect intervention.

Table 1 summarizes the performancetasks grouped under each posturalsystem for the BESTest. The entireBEStest with scoring, examiner, andpatient instructions is presented inthe eAppendix (available at: www.ptjournal.org). The performance tasksare grouped to reveal function ordysfunction of a particular systemunderlying balance control (see re-views by Horak and colleagues8–11).Here, we briefly summarize the roleof these systems in balance controland how each task item is related toits system:

I. Biomechanical Constraints: Bio-mechanical constraints for standingbalance include the quality of thebase of foot support (item 1), geo-metric postural alignment (item 2),

functional ankle and hip strength(force-generating capacity) for stand-ing (items 3 and 4), and ability to risefrom the floor to a standing position(item 5).39

II. Stability Limits/Verticality: Thissystem includes items for an internalrepresentation of how far the bodycan move over its base of supportbefore changing the support or los-ing balance, as well as an internalperception of postural vertical.40,41

The ability to lean as far as possiblein a sitting position with eyes closed(item 6) provides a measure of laterallimits of stability in a sitting posture,and the ability to realign the trunkand head back to perceived vertical(item 6) provides a measure of inter-nal representation of gravity. Theability to reach maximally forwardand laterally while standing (items 7and 8) represents the functional lim-its of stability, although this may not

necessarily be correlated with howfar a person can lean the body’s cen-ter of mass when not reaching.43,44

III. Anticipatory Postural Adjust-ments: This system includes tasksthat require an active movement ofthe body’s center of mass in antici-pation of a postural transition fromone body position to another. Forexample, we include the transitionsfrom a sitting to a standing position45

(item 9), from normal stance tostance on toes45 (item 10), and from2-legged- to 1-legged stance46 (item11). Item 12 involves repetitiveweight shifting from leg to leg inanticipation of tapping a forefoot ona stool, and item 13 involves antici-patory postural adjustments prior torapid, bilateral arm raises with aweight.47,48

IV. Postural Responses: Reactivepostural responses include both in-place and compensatory stepping re-sponses to an external perturbationinduced by the examiner’s hands us-ing the unique “push and release”technique.49 To induce an automaticpostural response with the patient’sfeet in place (ankle or hip strategy),the tester pushes isometrically againsteither the front (item 14) or back(item 15) of the patient’s shouldersuntil either the toes or the heels justbegin to raise without changing theinitial position of the body’s centerof mass over the feet before suddenlyletting go of the push. To induce com-pensatory stepping responses, thetester requires a forward (item 16) orbackward (item 17) or lateral (item18) lean of the patient’s center ofmass over the base of foot supportprior to release of pressure, requir-ing a fast, automatic step to recoverequilibrium.49,50

V. Sensory Orientation: This systemidentifies any increase in body swayduring stance associated with alter-ing visual or surface somatosensoryinformation for control of standing

Figure 1.Model summarizing systems underlying postural control corresponding to sections ofthe Balance Evaluation Systems Test (BESTest).



balance. Item 19 is the modified Clin-ical Test of Sensory Integration forBalance51 (CITSIB), and item 20 in-volves standing on a 10-degree,toes-up incline with eyes closed.

VI. Stability in Gait: This system in-cludes evaluation of balance duringgait (item 21) and when balance ischallenged during gait by changinggait speed52 (item 22), by head rota-tions53 (item 23), by pivot turns(item 24), and by stepping over ob-stacles54 (item 25). This section alsoincludes the Timed “Get Up & Go”Test, which evaluates how fast a pa-tient can sequence rising from achair, walking 3 m, turning, and sit-ting back down again without (item26) and with (item 27) a secondarycognitive task to challenge the pa-tient’s attention.55

Although several separate neural sys-tems underlie control of balance,

each task may involve more than onesystem that interacts with others. Forexample, the task of tapping alter-nate feet onto a stair (item 12) isplaced in the “Anticipatory PosturalAdjustments” system because it re-quires adequate anticipatory pos-tural weight shifting from one leg tothe other. However, it also requiresan adequate base of support andstrength in the hip abductors (“Bio-mechanical Constraints” system). In-teractions among systems can beseen by how a single pathology, suchas abnormal vestibular function, willlikely affect several tasks, such as theability to stand on foam with eyesclosed (item 19 in the “Sensory Ori-entation” system) and the ability torotate the head while walking (item23 in the “Stability in Gait” system).Future studies are needed to deter-mine the extent to which posturalsystem problems cluster, such that

disorders in each system can be dif-ferentiated in the clinic.

The purpose of this article is topresent the BESTest, with its theoret-ical framework and its first interraterreliability and concurrent validityanalysis. This is the first step in max-imizing the psychometric propertiesof this new balance assessment tool.

MethodDevelopment of the BESTestThe conceptual framework for de-veloping a balance assessment toolthat separates control of balance intoits underlying systems is based onthe scientific literature about labora-tory measures of postural disordersin elderly people and in people withneurological disorders.8–11 The prin-ciple of having physical therapistsevaluate 6 subcomponent systemsunderlying balance function initiallywas suggested as a qualitative assess-

Table 1.Summary of Balance Evaluation Systems Test (BESTest) Items Under Each System Categorya

I. BiomechanicalConstraints

II. StabilityLimits/Verticality

III. AnticipatoryPostural

AdjustmentsIV. PosturalResponses

V. SensoryOrientation VI. Stability in Gait

1. Base of support 6. Sittingverticality (leftand right) andlateral lean (leftand right)

9. Sit to stand 14. In-placeresponse, forward

19. Sensory integrationfor balance(modified CTSIB)

Stance on firmsurface, EO

Stance on firmsurface,EC

Stance on foam, EOStance on foam, EC

21. Gait, levelsurface

2. CoM alignment 7. Functionalreach forward

10. Rise to toes 15. In-placeresponse,backward

22. Change in gaitspeed

3. Ankle strengthand ROM

8. Functionalreach lateral(left and right)

11. Stand on oneleg (left andright)

16. Compensatorysteppingcorrection,forward

23. Walk with headturns, horizontal

4. Hip/trunk lateralstrength

12. Alternate stairtouching

17. Compensatorysteppingcorrection,backward

20. Incline, EC 24. Walk with pivotturns

5. Sit on floor andstand up

13. Standing armraise

18. Compensatorysteppingcorrection, lateral(left and right)

25. Step overobstacles

26. Timed “Get Up& Go” Test

27. Timed “Get Up& Go” Test withdual task

a CoM�center of mass, ROM�range of motion, CTSIB�Clinical Test of Sensory Integration for Balance, EO�eyes open, EC�eyes closed.



ment by Horak and Shumway-Cookin their continuing medical educa-tion courses between 1990 and1999.15–17,19,56,57 After Horak andFrank developed the BESTest, thou-sands of experienced physical ther-apy clinicians contributed to contin-ued development of the BESTest byproviding feedback about clarity,sensitivity, and practicality of itemsacross 38 continuing educationworkshops delivered by Horak be-tween 1999 and 2005. Following 2days of didactic and observationaltraining in the test, therapists in theworkshops practiced performanceof the test on each other and pro-vided critical feedback to improvethe clarity and specificity of instruc-tions to patients and therapists.Some of the balance tasks in the testhave been borrowed from currentassessment tools, although they arenow placed within our theoreticalframework and the therapist and pa-tient instructions, and most of therating scales have been modified toimprove consistency and reliability(Tab. 2). This is the first balance as-sessment tool to include a clinicalmethod for assessing postural re-sponses to external perturbations (sec-tion IV) and verticality (section II).

The BESTest consists of 27 tasks,with some items consisting of 2 of 4subitems (eg, for left and right sides),for a total of 36 items. Each itemis scored on a 4-level, ordinal scalefrom 0 (worst performance) to 3(best performance). Scores for thetotal test, as well as for each section,are provided as a percentage of totalpoints. Specific patient and rating in-structions and stopwatch and rulervalues are used to improve reliability(see the eAppendix for the full test).

Session 1: Raters and SubjectsTo evaluate the interrater reliabilityand internal consistency of the orig-inal version of the BESTest (currentsections II–VI), we recruited 12 am-bulatory adults with a wide range ofbalance function. Subjects were re-cruited as a sample of conveniencefrom individuals who previously hadparticipated in research studies onbalance and postural control. Nosubjects had completed the BESTestprior to the first session. However,subjects may have completed spe-cific items that were adapted fromother clinical tests such as the Dy-namic Gait Index. For this session,we included 3 subjects with PD, 5subjects with vestibular dysfunction

(3 with bilateral loss, 2 with unilat-eral loss), 1 subject with peripheralneuropathy and a total hip arthro-plasty, and 3 subjects who werehealthy (controls) (Tab. 3). All sub-jects met the following inclusioncriteria: (1) ability to follow 3-stepcommands, (2) ability to provide in-formed consent, (3) ability to ambu-late 6 m (20 ft) without human assis-tance, and (4) ability to tolerate thebalance tasks without excessive fa-tigue. Subjects were provided shortrest breaks as needed. The subjects(5 female, 7 male) ranged in age from50 to 80 years (X�63, SD�10). De-scriptive information for the subjectswho completed the BESTest is listedin Table 3. None of the subjects usedan assistive device during the testing.

The 9 raters consisted of a conve-nience sample of 6 physical thera-pists from various practice settingsand 3 Doctor of Physical Therapystudents from Pacific University(mean age�33.1 years, SD�4.7; 3male, 6 female; Tab. 3). Physical ther-apists were included if they had avalid Oregon physical therapist li-cense, and physical therapist stu-dents were included if they had com-pleted the relevant course work in

Table 2.Balance Tasks in the Balance Evaluation Systems Test (BESTest) That Have Been Borrowed From Existing Clinical Testsa

Clinical Test

FunctionalReach Test

Fregly Single-LimbStance Test

Berg BalanceScale CTSIB Dynamic Gait Index

Timed “Up &Go” Test

BESTestItem

7. Functional reachforward

11. Stand on one leg,right

12. Alternate stairtouching

19. Sensory integrationfor balance, stanceon firm surface, EO

21. Gait, level surface 26. Timed “Get Up &Go” Test

8. Functional reachlateral, right

11. Stand on one leg,left

19. Sensory integrationfor balance, stanceon firm surface, EC

22. Change in gaitspeed

8. Functional reachlateral, left

19. Sensory integrationfor balance, stanceon foam, EO

23. Walk with headturns, horizontal

19. Sensory integrationfor balance, stanceon foam, EC

24. Walk with pivotturns

25. Step overobstacles

a CTSIB�Clinical Test of Sensory Interaction on Balance, EO�eyes open, EC�eyes closed.



relation to the evaluation and treat-ment of balance disorders.

Session 2: Raters and SubjectsAfter initial analysis of the first reli-ability data, a second testing ses-sion 18 months later evaluated theinterrater reliability of a newly devel-oped section I (“Biomechanical Con-straints”) and a revised section VI(“Stability in Gait”). Section VI wasrevised due to a low intraclass corre-lation coefficient (ICC [2,1]�.54) ob-tained in the first session. The goalsof this second testing session wereto improve the reliability of sectionVI by modifying the criteria for scor-ing and requiring raters to view sub-jects from the front or back whilewalking and to add section I onbiomechanical constraints affectingpostural control. Testing session 2involved 11 raters, including 3 raters

from the first session (denoted byasterisks in Tab. 3). No students wereincluded, although 2 raters were PhDresearchers in human balance disor-ders without any physical therapytraining or experience (Tab. 3).Eleven subjects, including 4 subjectsfrom the first session, were adminis-tered 2 sections of the BESTest. As inthe first session, subjects were a sam-ple of convenience recruited fromindividuals who had previously par-ticipated in laboratory studies butwho had no experience with theBESTest. Subjects in session 2 metthe same inclusion criteria as in ses-sion 1. The subjects consisted of 6subjects who were healthy (con-trols), 1 subject with unilateral ves-tibular loss, 1 subject with bilateralvestibular loss, 2 subjects with PD,and 1 subject with both peripheralneuropathy and bilateral hip arthro-

plasty. The subjects (5 female, 6male) ranged in age from 67 to 88years (X�75, SD�7.6).

The data and analysis from sections Ithrough IV (current sections II–V) ofsession I and the new section I andrevised section VI from session 2 arepresented in this article. For bothsessions, each subject completed aninformed consent statement accord-ing to the Declaration of Helsinki.

ProcedureAll raters were provided with theBESTest and written instructions foradministering the test approximately1 week prior to the session. On theday of the study, the raters partici-pated in a 45-minute training sessionwith one of the developers of theBESTest (FBH). For training raters,each item of the BESTest was dem-

Table 3.Descriptive Information on the Raters and Subjectsa

Descriptive information on the Raters Using the BESTest

RaterNo.

Years ofPractice

ClinicalSetting

OrthopedicExperience

BalanceExperience

NeurologicExperience

Session 1: BESTestSections II–VI

1 0 Student 0 0 0

2 0 Student 0 0 0

3 * 0 Student 0 0 0

4 4 Research 1 0 4

5 4 OP orthopedics 4 0 0

6 5 IP acute care 5 0 0

7 12 OP neurology 0 12 12

8 13 Faculty 0 10 10

9 * 19 Research 3 13 19

Session 2: BESTestSection I andRevised Section VI

1 0 Research 0 0 0

2 0 Research 0 0 0

3 * 1 OP neurology 0 1 1

4 4 Research 0 4 4

5 14 Home care 0 11 11

6 19 OP orthopedics 16 15 15

7* 20 Faculty 3 14 20


9 22 Faculty 1 15 22


(Continued)



onstrated on a subject who did notparticipate in the reliability study,and the rating criteria were dis-cussed. The raters were allowed toask questions regarding the scoringof the test. However, the raters wereinstructed to rate each outcome withno assistance or discussion with theother raters. The BESTest took 20 to30 minutes to administer.

During the experimental sessions,the raters were asked to concur-rently rate each of the subjects. Inboth sessions 1 and 2, one of theauthors (FBH), who was not one ofthe raters, administered the BESTest

once for each subject while theraters observed. The raters were al-lowed to position themselves aroundthe area where the subjects wereperforming the test and to moveabout as needed in order to opti-mally view the subjects’ perfor-mance for recording the outcome.Only one opportunity was providedto view the performance of eachtest item. If a rater missed the per-formance of an item, the item wasrepeated (3 items for session 1 and 1item for session 2), and all of theraters scored the second perfor-mance for consistency. Raters wereprovided with separate scoring sheets

for each subject and did not discussscoring among subjects. The raterswere instructed to rate each out-come independently, with no assis-tance from or discussion with theother raters. The diagnoses of thesubjects who completed the BESTestwere masked from the raters.

To begin to describe concurrentvalidity, subjects completed the Acti-vities-specific Balance Confidence(ABC) Scale.58 The ABC Scale quan-tifies how confident a person feelsthat he or she will not lose balancewhile performing 16 activities ofdaily living. The ABC Scale has dem-

Table 3Continued

Descriptive Information on the Subjects Completing the BESTest

SubjectNo. Age (y) Sex Diagnosis

ABCScale Falls

BESTestTotal Score

Session 1: BESTestSections I–VI

1 56 F Control 99 0 95.83

2* 64 M Control 99 0 86.46

3* 77 M Control 95 0 89.58

4 56 F UVL 66 0 79.17

5 80 M UVL 78 0 61.46

6 50 F BVL 55 0 71.88

7 57 F BVL 64 1 83.33

8 64 M BVL 72 0 88.54

9 62 M PD 57 0 68.75

10* 65 M PD 70 0 78.13

11 75 M PD 53 0 63.54

12* 75 F PNP, THA 73 0 76.04

Session 2: BESTestSection I andRevised Section VI

1 53 M UVL 94 0 NA

2 66 F Control 100 0 NA

3* 67 M Control 93 0 NA


5* 79 M Control 93 0 NA


7 83 M Control 93 0 NA

8 88 F BVL 57 1 NA

9* 68 M PD 73 0 NA

10 69 M PD 83 0 NA

11* 77 F PNP, THA 74 0 NA

a BESTest�Balance Evaluation Systems Test. ABC Scale�Activities-specific Balance Confidence Scale, OP�outpatient, IP�inpatient, F�female, M�male,UVL�unilateral vestibular loss, BVL�bilateral vestibular loss, PD�Parkinson disease, PNP�peripheral neuropathy, THA�total hip arthroplasty, NA�notapplicable. Asterisk indicates participation in both sessions.



onstrated test-retest reliability(r�.92).59 Scores on the ABC Scalerange from 0, indicating no confi-dence, to 100, indicating completeconfidence in the person’s ability toperform the task without losing bal-ance. Scores on the ABC Scale havebeen correlated with ratings of olderadults’ level of community function.60

Data AnalysisInterrater agreement for individualBESTest items was determined usingthe Kendall coefficient of concor-dance for ordinal data.61 Concurrentvalidity was assessed by analyzingthe correlation of the BESTest totaland subsection scores of the raterwith the most exposure to the BEST-est (DMW) with the ABC Scalescores using the Spearman correla-tion coefficient. Coefficients of .00to .25 were interpreted to indicatelittle to no relationship, .25 to .50as a fair relationship, .50 to .75 as amoderate to good relationship, andabove .75 as a strong relation-ship.1,2,4,5,62 A Mann-Whitney U testwas used on the ranking of BESTesttotal scores (of the rater with themost exposure to the BESTest)among subjects to determinewhether the 3 controls scored betterthan the 7 subjects with balanceproblems.

ResultsInterrater ReliabilityInterrater reliability statistics forBESTest total and subsection scoresare presented in Table 4. The inter-rater reliability of the BESTest totalscores was excellent, with an ICC(2,1) of .91. Subsection ICCs rangedfrom .79 to .96, and Kendall coeffi-cients ranged from .79 to .95, indi-cating good to excellent reliability.Reliability statistics for individualBESTest items are presented in Fig-ure 2. Individual items demonstratedKendall coefficients ranging from .46to 1.00. Items based on stopwatchtime, such as items in section V(“Sensory Orientation”), tended to

show the highest concordance,whereas judgments of alignment, an-kle strength, and sitting limits ofstability and verticality tended toshow the lowest concordance. Only3 items could not have concordancemeasured accurately because of lim-ited variability among subjects (de-noted by asterisks in Fig. 2). All ratersscored all subjects as excellent(score of 3) on standing arm raise,and they scored the majority of sub-jects as excellent on the alternatestair touch (92%) and stance witheyes open (98%).

The ICCs for the BESTest total scoreswere .94 among the 3 students and.87 among the 6 therapists. The ICCsfor each item within section VI(“Stability in Gait”) improved for thesecond interrater testing sessioncompared with the first session, byinstructing raters to view the sub-jects’ gait from the front or backrather than from the side. The ICCsfor the BESTest total scores for itemsin section VI increased from .54 to.88, with the range of Kendall coef-ficients for individual items of .51 to.72 for the first interrater testing ses-sion increasing to a range of .62 to.90 for the second interrater testingsession.

Test PerformanceThe subjects showed a wide range ofvariability on their performance ofthe test (Fig. 3). Figure 3 presents themedian and interquartile ranges ofBESTest total scores (expressed aspercentages) across diagnostic cate-gories. Median scores of all subjectsranged from 65% to 95%, with con-trol subjects clustered at the highend and subjects with PD clusteredat the low end. The Mann-Whitney Utest showed that control subjectsscored significantly higher (better)than the subjects with balance prob-lems (P�.036).

Consistent with our theoretical con-struct, the scores for each BESTestsection by diagnostic subgroup(Tab. 5) show that the subjects withunilateral vestibular loss scored theworst in section V (“Sensory Orien-tation”) (60%), whereas the sub-jects with PD scored the worst insection IV (“Postural Responses”)(50%). The 1 subject with neuropa-thy scored the worst on section III(“Anticipatory Postural Adjustments”).Although this score was similar tothat of the subjects with unilateralvestibular loss (67% versus 69%), thesubject with neuropathy could bedistinguished by a much higher

Table 4.Interrater Reliability Statistics for Balance Evaluation Systems Test (BESTest) Sectionand Total Scoresa

BESTest SectionICC,

Mean (95% CI)

Kendall Coefficient ofConcordance for

Ordinal Measures,Mean (95% CI)

Section I. Biomechanical Constraints .80 (.63–.93) .79 (.73–.85)

Section II. Stability Limits/Verticality .79 (.63–.92) .86 (.84–.88)

Section III. Anticipatory PosturalAdjustments

.92 (.85–.97) .92 (.91–.93)

Section IV. Postural Responses .92 (.85–.97) .91 (.90–.92)

Section V. Sensory Orientation .96 (.92–.99) .95 (.947–.953)

Section VI. Stability Gait .88 (.76–.96) .93 (.90–.96)

Total .91 (.83–.97)

a ICC�intraclass correlation coefficient, CI�confidence interval.



score on section V (“Sensory Orien-tation”) (93% versus 60%) and ahigher BESTest total score (79% ver-sus 73%).

The most difficult items for our sub-jects were: single-limb stance, stanceon foam with eyes closed, Timed“Get Up & Go” Test with a cognitivetask, walk with horizontal headturns, backward in-place postural re-sponses, and standing hip strength.In one item (standing arm raise),all subjects had perfect scores; theother least-difficult items includedstance with eyes open, alternate stairtouch, sitting verticality and leans,and stance with eyes closed. Sortedby difficulty, the mean score (SD)

and the frequency of how often ascore was given for an individualitem are summarized in Table 6. Vari-ability among subjects and ratersprovided a wide range of scoresacross BESTest items.

Concurrent Validity WithABC ScaleThe BESTest total scores correlatedsignificantly with each subject’s bal-ance confidence, as measured by thesubject’s average ABC Scale score(r�.685, r2�.47, P�.05; Fig. 4). TheABC Scale scores demonstrated mod-erate correlation with the BESTestsection scores of the BESTest(r�.41–.78). Section II (“StabilityLimits/Verticality”) scores had the

best correlation with the ABC Scalescores (r�.78), and Section III (“An-ticipatory Postural Adjustments”)scores had the worst correlation(r�.41).

Discussion and ConclusionsThis study presents a new clinicalbalance assessment tool that is thefirst tool aimed at distinguishing theunderlying systems that may be con-tributing to balance problems in in-dividual patients. By distinguishingwhich systems underlying balancecontrol are affected, this is the firstclinical balance assessment tool tohelp direct rehabilitation of peoplewith balance disorders. The most im-portant contribution of the BESTest

Figure 2.Kendall coefficient of concordance scores for individual items of the Balance Evaluation Systems Test (BESTest). Error bars indicate95% confidence interval. Asterisk indicates Kendall coefficient of concordance unable to be calculated accurately due to lack ofvariance in the data. EO�eyes open, EC�eyes closed.



is to provide a conceptual frame-work around which to evaluate andtreat patients with different types ofbalance problems.

Most existing clinical balance testsare directed at predicting fall risk orwhether a balance problem exists,rather than what type of balanceproblem exists.1–6 Although thesetests have proven valid in predictingthe likelihood of future falls, withsensitivity and specificity values of80% to 90%, the test results do nothelp therapists direct treatment.63–65

Lord et al1 developed a different typeof test, directed at identifying physi-ological impairments that could af-fect balance, such as impaired pro-prioception, visual function, orreaction time delays. Although thetest is helpful for understanding thephysiological reasons for balanceproblems, it is not apparent how totranslate many of the impairmentsinto specific balance exercise pro-grams. Identification of impairmentsmay help to identify the pathology,such as peripheral neuropathy orvestibular loss, that may be responsi-ble for the balance problem. How-ever, therapeutic exercise is not bestdesigned based on pathology, be-cause the functional ability of eachpatient is multifactorial and dependsnot only on the patient’s pathologybut also on the patient’s compensa-tion, experience, motivation, priorand concurrent pathologies, age, andso on.

It is especially critical, however, tostop conceptualizing balance as asingle system so that treatment canbe more specific than generalized“balance training” for a generalized“balance problem.” There is little ev-idence of carryover from learningone motor task to a different motortask, so practicing grapevine step-ping in balance training is unlikely toimprove functional limits of stability,postural responses to perturbations,or the ability to use vestibular in-

formation for balance. If a patientshows difficulty on a particular sec-tion of the BESTest, the therapistshould not limit therapy to practic-ing the specific tasks that were diffi-cult for the patient but should aimtherapy at the underlying systemdeficit.66

If the BESTest is valid in supportingthe conceptual framework that bal-

ance function can be divided intoseparate underlying systems, wewould expect some patients to per-form poorly in different subcate-gories compared with other pa-tients. Even with our small sample ofsubjects, the 3 subjects with PDtended to perform poorly on items insection IV (“Postural Responses”),whereas the 3 subjects with vestibu-lar loss performed poorly on items

Table 5.Percentage Score in Each Balance Evaluation Systems Test (BESTest) Section andTotal Score in Session 1 by Diagnostic Group

DiagnosticGroup Section II Section III Section IV Section V Section VI Total

Control (n�3) 100 81 88 91 89 94

BVL (n�3) 83 76 83 78 84 85

PNP (n�1) 71 67 78 93 76 79

UVL (n�2) 75 69 69 60 74 73

PD (n�3) 76 72 50 71 78 73

a BVL�bilateral vestibular loss, PNP�peripheral neuropathy, UVL�unilateral vestibular loss,PD�Parkinson disease.

Figure 3.Median and interquartile range of Balance Evaluation Systems Test (BESTest) totalscores across diagnostic categories for sections II through VI in testing session 1. Notethe variation in scores among subjects tested. UVL�unilateral vestibular loss,BVL�bilateral vestibular loss, PD�Parkinson disease, PNP�peripheral neuropathy.



Table 6.Means, Standard Deviations, and Distribution of Balance Evaluation Systems Test (BESTest) Scores Within Each Balance ItemListed by Item Difficultya

Section Item Mean SD

Frequency

0 1 2 3 Total

III 11. Stand on one leg, right 1.13 0.98 29 51 11 17 108

III 11. Stand on one leg, left 1.18 1.01 29 52 9 18 108

V 19. Stance on foam, EC 1.30 1.01 18 63 3 24 108

VI 27. Timed “Get Up & Go” Test with dual task 1.33 1.06 31 29 31 18 109

VI 23. Walk with head turns, horizontal 1.40 1.03 21 48 17 24 110

IV 15. In-place response, backward 1.57 0.79 9 37 54 8 108

I 4. Hip/trunk lateral strength 1.59 1.17 31 13 36 30 110

VI 25. Step over obstacles 1.87 1.24 28 8 24 50 110

VI 24. Walk with pivot turns 1.95 0.95 2 46 18 44 110

I 3. Ankle strength and ROM 1.99 1.09 11 32 14 53 110

IV 17. Compensatory stepping correction, backward 2.18 0.62 2 7 68 31 108

I 5. Sit on floor and stand up 2.19 1.25 21 10 2 72 105

II 8. Functional reach lateral, left 2.22 0.41 0 0 84 24 108

IV 14. In-place response forward 2.26 0.66 1 8 65 34 108

V 19. Stance on foam, EO 2.26 1.03 9 19 15 65 108

III 10. Rise to toes 2.32 0.72 0 17 42 49 108

IV 16. Compensatory stepping correction, forward 2.32 0.71 2 8 52 46 108

II 8. Functional reach lateral, right 2.38 0.49 0 0 67 41 108

VI 21. Gait, level surface 2.39 0.78 2 14 33 61 110

IV 18. Compensatory stepping correction, lateral (left) 2.40 0.75 0 16 31 60 107

IV 18. Compensatory stepping correction, lateral (right) 2.43 0.86 0 27 9 72 108

I 1. Base of support 2.48 0.86 4 14 17 74 109

I 2. CoM alignment 2.49 0.81 4 10 24 72 110

V 20. Incline, EC 2.53 0.70 1 8 29 70 108

II 7. Functional reach forward 2.56 0.50 0 0 48 60 108

VI 26. Timed “Get Up & Go” Test 2.56 0.78 2 13 16 77 108

VI 22. Change in gait speed 2.60 0.79 0 21 2 87 110

II 6. Sitting lateral lean, right 2.61 0.61 0 7 24 77 108

II 6. Sitting lateral lean, left 2.65 0.50 0 1 32 75 108

II 6. Sitting verticality, left 2.66 0.51 0 2 29 77 108

V 19. Stance on firm surface, EC 2.66 0.75 0 18 1 89 108

III 9. Sit to stand 2.73 0.69 0 13 0 94 107

II 6. Sitting verticality, right 2.85 1.84 0 3 29 76 108

III 12. Alternate stair touching 2.92 0.28 0 0 9 99 108

V 19. Stance on firm surface, EO 2.98 0.13 0 0 1 107 108

III 13. Standing arm raise 3.00 0.00 0 0 0 108 108

a Frequency�how often each rating was provided for an individual item. Total�total number of ratings for each item (items have different totals due tomissing data). Means, standard deviations, and frequency for items 6–20 reported for session 1 and frequency for items 1–5 and 21–27 reported for session2. EO�eyes open, EC�eyes closed, ROM�range of motion, CoM�center of mass.



in section V (“Sensory Orientation”).Laboratory studies of postural re-sponses and the ability to maintainequilibrium in stance under differentsensory conditions in patients withPD, unilateral vestibular loss, or bi-lateral vestibular loss support thesetrends in our study.67–69 In contrastto the subjects with PD and vestibu-lar loss, the one subject with periph-eral neuropathy combined with bi-lateral total hip arthroplasty scoredworst on items in section III (“An-ticipatory Postural Adjustments”).Based on these differential results,therapists would direct the patientswith PD to practice compensatorystepping in response to perturba-tions,70 the patients with unilateralvestibular loss to practice balancingin conditions requiring use of theremaining vestibular information,66

and the patient with peripheral neu-ropathy to practice moving from onestable posture to another.62 Ofcourse, other patients with thesesame pathologies may show differ-ent profiles in the BESTest, depend-ing on their compensation strategies,which may affect their ability toovercome limitations from physio-logical constraints to perform a taskusing an alternative strategy.

Although the categories of systemsin the BESTest were selected fromcurrent, scientific understanding ofneurophysiological systems underly-ing postural control, the systems arequite interdependent. For example,constraints on the base of foot sup-port (item 1) will necessarily affectthe forward limits of postural stabil-ity in standing (item 7), and difficultyusing vestibular information to standon foam with eyes closed (item 19D)may make it difficult to perform headturns during gait (item 23). Further-more, the tasks selected to revealfunction of each of the 6 posturalsystems may not be ideal; some tasksare likely too easy to be discrimina-tory. For example, the standing armraise to look for anticipatory postural

adjustments (item 13) and stancewith eyes open to examine posturalsway (item 19) may only be sensitivein a laboratory, where surface reac-tive forces or body kinematics can bemeasured to detect physiologicallysignificant, but not clinically appar-ent, changes in postural control. Allof our subjects also scored a perfect3 on alternate stair touch (item 12),adapted from the Berg BalanceScale,62 but this may be a problemwith the excessively long time crite-ria (within 20 seconds) for doingonly 8 steps, so we recommend in-creasing the number of steps to 15 inorder to determine the number ofsteps completed per second. Furtherpsychometric testing on largegroups of patients with a variety of

balance problems will reveal whichitems naturally group together andmay suggest that some items shouldbe moved or eliminated or altered, oreven that a new system categoryshould be added (ie, cognitive inter-ference with balance performance).

With an ICC of .91 for BESTest totalscores, the interrater reliability forthe BESTest is excellent71 and justas good, or better than, the cur-rent, shorter balance assessmentbatteries (Berg Balance Scale: ICC�.9872; Tinetti Mobility Assessment:ICC�.75–1.042). Subsections of theBESTest adapted from established testsin the literature also show reliabilitysimilar to or better than that previ-ously reported: Functional Reach Test

Figure 4.Correlation between subjects’ Activities-specific Balance Confidence (ABC) Scale meanscores and their Balance Evaluation Systems Test (BESTest) total scores (from testingsession 1 raters’ median scores).



ICC�.9873 compared with BESTestsection II ICC�.79; CTSIB ICC�.7474 compared with BESTest sectionV ICC�.96; Dynamic Gait Indexkappa�.642 and Timed “Get Up &Go” Test ICC�.997 compared withBESTest section VI ICC�.88. The in-terrater reliability of each section ofthe BESTest is sufficiently strong toallow therapists to use an individualsection if they are short on time orwant to direct a balance test at aspecific postural system.75 An abbre-viated test would be helpful becausethe BESTest takes about 30 minutesto carry out, even by an experiencedtherapist. Future studies are neededto identify redundant and insensitiveitems and to eliminate unnecessaryitems that do not add value to thetest.

Inexperienced raters, without phys-ical therapy experience, were able tolearn how to score the BESTest withprior review and 45 minutes of in-struction with demonstration. Thisunfamiliarity may have caused ratersto be unsure of how to score a par-ticular item or to make an errorwhen recording a score. The reliabil-ity of Peabody Motor DevelopmentalScales-2 scores has been shown toincrease as familiarity with the testincreased.76 Because some of theitems are novel and required specifichand positions and instructions, ac-tual demonstration and training maybe necessary for excellent interraterreliability, as well as for safety. Spe-cifically, the push and release tech-nique to elicit automatic postural re-sponses by suddenly releasing thesubjects’ leans requires observationand practice with at least video dem-onstration. Because the compensa-tory stepping postural responsesnecessarily required to move thebody’s center of mass beyond thelimits of the base of foot support,these items also are the most danger-ous to test in patients with balancedisorders and, therefore, require spe-cial training. In some cases, subjects

who are judged to be prone to a fallif attempting these items should au-tomatically receive a score of 0 ornot be tested in order to avoid a fall.Some scores, such as those for sec-tion IV (“Postural Responses”), mayhave been even more reliable if theraters also were physically perform-ing the BESTest, although otherscores, such as those for functionalreach (items 7–9), were likely betterbecause subjects could be viewedfrom a distance, without standby as-sistance for safety in our study. Inthis study, we found that it is impor-tant for raters to stand in front or inback of subjects, rather than parallelwith them, while they are walking initems for section VI (“Stability inGait”) in order to view potential lat-eral postural instability during gait.To improve reliability, we have sincedeveloped an educational DVD totrain therapists how to administerand score the BESTest.*

The strong agreement between theBESTest total score and subjects’ rat-ing of their balance confidence inthe ABC Scale suggests that the BEST-est measures aspects of balance func-tionally relevant to patients. TheABC Scale has been shown to be re-lated to patients’ actual unwilling-ness to engage in activities in thecommunity due fear of falling.59

However, treatments cannot be de-signed based solely on the ABCScale, and a current study is investi-gating the relationship between theBESTest and the Berg Balance Scaleand prospective falls in patients witha wide range of pathologies andabilities.

LimitationsThis study had several limitations. Itis possible that other systems impor-

tant for balance control are missingfrom the test and only the last item isrelated to cognitive constraints onbalance, and this may be inadequate.Whether or not the sections of theBESTest accurately detect dissociablebalance deficits remains to be inves-tigated to establish its construct va-lidity. How well section III (“PosturalResponses”) and section IV (“Sen-sory Orientation”) are related to sim-ilar measures using computerizedposturography is unknown. SectionsI and II should be revised to im-prove their test-retest reliability. Inaddition, the test is quite long, suchthat future clinimetric studies needto identify redundant, insensitiveitems for a more efficient clinicaltool. We also do not know how sen-sitive the BESTest is to change withintervention.

Further psychometric testing is war-ranted for the BESTest to establish itsconstruct and concurrent validity,sensitivity and specificity, and abilityto direct effective treatment for peo-ple with balance disorders. The scaleis quantitative, and scoring is repro-ducible both for the test as a wholeand for its subsections, as demon-strated by agreement among raterswith varying experience. The BEST-est appears to be testing functionallyrelevant aspects of balance controlas seen by the agreement with sub-jects’ self-reported balance confi-dence. However, success of theBESTest will depend on how useful itis in assisting therapists to organizetheir systematic assessment of bal-ance disorders to develop specifictreatments based on each individu-al’s balance constraints.

All authors provided concept/idea/projectdesign and writing. Dr Horak and Dr Wrisleyprovided data collection and analysis,project management, fund procurement,and subjects. Dr Horak provided facilities/equipment, institutional liaisons, and clericalsupport. Dr Horak and Dr Frank provided

* The BESTest Training DVD is distributedthrough Oregon Health and Science Universi-ty’s Technology and Research CollaborationsOffice and is available via a nonexclusive li-cense. See: http://www.ohsu.edu/tech-transfer/portal/technology.php?technology_id�217191.



consultation (including review of manuscriptbefore submission).

The authors thank Larry Meyer and TrentThompkins for collecting data on the firstinterrater reliability study as part of theirDoctor of Physical Therapy thesis, as well asall of the subjects and raters who partici-pated in this study. The authors also areindebted to the physical therapists whoprovided helpful criticisms of early versionsof the test in continuing education work-shops by Dr Horak. Statistical support fromDr George Knafl and Dawn Peters also isappreciated.

This work was supported by the NationalInstitute on Aging grant R0-1 AG006457.

Poster presentations of this research weregiven at the Combined Sections Meetingsof the American Physical Therapy Associa-tion; February 4–8, 2004; Nashville, Tennes-see; and February 23–27, 2005; New Or-leans, Louisiana.

This article was received March 10, 2008, andwas accepted January 30, 2009.

DOI: 10.2522/ptj.20080071

References1 Lord SR, Clark RD, Webster IW. Physiolog-

ical factors associated with falls in an el-derly population. J Am Geriatr Soc. 1991;39:1194–1200.

2 Podsiadlo D, Richardson S. The timed “Up& Go”: a test of basic functional mobilityfor frail elderly persons. J Am Geriatr Soc.1991;39:142–148.

3 Berg KO, Maki BE, Williams JI, et al. Clin-ical and laboratory measures of posturalbalance in an elderly population. ArchPhys Med Rehabil. 1992;73:1073–1080.

4 Shumway-Cook A, Baldwin M, Polissar NL,Gruber W. Predicting the probability forfalls in community-dwelling older adults.Phys Ther. 1997;77:812–819.

5 Shumway-Cook A, Brauer S, WoollacottMH. Predicting the probability for falls incommunity-dwelling older adults usingthe Timed “Up & Go” Test. Phys Ther.2000;80:896–903.

6 Tinetti ME, Baker DI, McAvay G, et al. Amultifactorial intervention to reduce therisk of falling among elderly people livingin the community. N Engl J Med. 1994;331:821–827.

7 Whitney SL, Poole J, Cass S. A review ofbalance instruments for older adults. Am JOccup Ther. 1998;52:666–671.

8 Horak FB. Postural orientation and equilib-rium: what do we need to know aboutneural control of balance to prevent falls?Age Ageing. 2006;35(Suppl 2):ii7–ii11.

9 Horak FB, Macpherson JM. Postural orien-tation and equilibrium. In: Smith JL, ed.Handbook of Physiology: Section 12—Exercise: Regulation and Integration ofMultiple Systems. New York, NY: OxfordUniversity Press; 1996:255–292.

10 Macpherson J, Horak FB. Neural control ofposture. In: Kandel E, Schwartz J, Jessell T,eds. Principles of Neural Science. 5th ed.New York, NY: Elsevier; in press.

11 Horak FB, Shupert CL, Mirka A. Compo-nents of postural dyscontrol in the elderly:a review. Neurobiol Aging. 1989;10:727–738.

12 Woollacott MH, Shumway-Cook A. Atten-tion and the control of posture and gait: areview of an emerging area of research.Gait Posture. 2002;16:1–14.

13 Nutt J, Horak FB. Gait and balance disor-ders. In: Asbury AK, McKhann GM, Mc-Donald WI, et al, eds. Diseases of theNervous System: Clinical Neuroscienceand Therapeutic Principles. 3rd ed. Cam-bridge, United Kingdom: Cambridge Uni-versity Press; 2002:581–591.

14 Bernstein NA. The Co-ordination andRegulation of Movements. Oxford, NY:Pergamon Press; 1967.

15 Horak FB, Shumway-Cook A. Clinical im-plications of posture control research. In:Duncan P, ed. Balance: Proceedings ofthe APTA Forum. Alexandria, VA: Ameri-can Physical Therapy Association; 1990:105–111.

16 Horak FB. Effects of neurological disorderson postural movement strategies in theelderly. In: Vellas B, Toupet M, RubensteinL, et al, eds. Falls, Balance, and Gait Dis-orders in the Elderly. Paris, France:Elsevier Science Publishers; 1992:137–151.

17 Horak FB. Clinical measurement of pos-tural control in adults. Phys Ther. 1987;67:1881–1885.

18 Horak FB, Henry SM, Shumway-Cook A.Postural perturbations: new insights fortreatment of balance disorders. Phys Ther.1997;77:517–533.

19 Horak FB. Clinical assessment of balancedisorders Gait Posture. 1997;6:76–84.

20 Horak FB, Frank J. Three separate posturalsystems affected in parkinsonism. In: Stu-art DG, Gurfunkel VS, Wiesendanger M,eds. Motor Control VII. Tucson, AZ: MotorControl Press; 1996:343–346.

21 Robinovitch SN, Heller B, Lui A, Cortez J.Effect of strength and speed of torque de-velopment on balance recovery with theankle strategy. J Neurophysiol. 2002;88:613–620.

22 Jacobs JV, Dimitrova DM, Nutt JG, HorakFB. Can stooped posture explain multidi-rectional postural instability in patientswith Parkinson’s disease? Exp Brain Res.2005;166:78–88.

23 Mancini M, Rocchi L, Horak FB, Chiari L.Effects of Parkinson’s disease and levo-dopa on functional limits of stability.Clin Biomech (Bristol, Avon). 2008;23:450–458.

24 Bisdorff AR, Anastasopoulos D, BronsteinAM, Gresty MA. Subjective postural verti-cal in peripheral and central vestibulardisorders. Acta Otolaryngol Suppl. 1995;520(Pt 1):68–71.

25 Burleigh-Jacobs A, Horak FB, Nutt JG,Obeso JA. Step initiation in Parkinson’sdisease: influence of levodopa and exter-nal sensory triggers. Mov Disord. 1997;12:206–215.

26 Horak FB, Esselman P, Anderson ME,Lynch MK. The effects of movement ve-locity, mass displaced, and task certaintyon associated postural adjustments madeby normal and hemiplegic individuals.J Neurol Neurosurg Psychiatry. 1984;47:1020–1028.

27 Jacobs JV, Horak FB. Cortical control ofpostural responses. J Neural Transm.2007;114:1339–1348.

28 Horak FB, Diener HC. Cerebellar controlof postural scaling and central set instance. J Neurophysiol. 1994;72:479–493.

29 Cameron MH, Horak FB, Herndon RR,Bourdette D. Imbalance in multiple sclero-sis: a result of slowed spinal somatosen-sory conduction. Somatosens Mot Res.2008;25:113–122.

30 Inglis JT, Horak FB, Shupert CL, Jones-Rycewicz C. The importance of somato-sensory information in triggering andscaling automatic postural responses inhumans. Exp Brain Res. 1994;101:159–164.

31 Horak FB. Adaptation of automatic pos-tural responses. In: Bloedel J, Ebner TJ,Wise SP, eds. Acquisition of Motor Behav-ior in Vertebrates. Cambridge, MA: MITPress; 1996:57–85.

32 Peterka RJ, Loughlin PJ. Dynamic regula-tion of sensorimotor integration in humanpostural control. J Neurophysiol. 2004;91:410–423.

33 Speers RA, Kuo AD, Horak FB. Contribu-tions of altered sensation and feedback re-sponses to changes in coordination of pos-tural control due to aging. Gait Posture.2002;16:20–30.

34 Yang JF, Winter DA, Wells RP. Posturaldynamics of walking in humans. Biol Cy-bern. 1990;62:321–330.

35 King LA, Horak FB. Lateral stepping forpostural correction in Parkinson’s disease.Arch Phys Med Rehabil. 2008;89:492–499.

36 Nashner LM, Black FO, Wall C III. Adapta-tion to altered support and visual condi-tions during stance: patients with vestibu-lar deficits. J Neurosci. 1982;2:536–544.

37 Horak FB. Role of the vestibular system inpostural control. In: Herdman SJ, ed. Ves-tibular Rehabilitation. 3rd ed. Philadel-phia, PA: FA Davis Co; 2007:25–51.

38 O’Sullivan SB. Assessment of motor func-tion. In: O’Sullivan SB, Schmitz TJ, eds.Physical Rehabilitation: Assessment andTreatment. 4th ed. Philadelphia, PA: FADavis Co; 2001:191–197.

39 Hayes KC. Biomechanics of postural con-trol. Exerc Sport Sci Rev. 1982;10:363–391.

40 McCollum G, Leen TL. Form and explora-tion of mechanical stability limits in erectstance. J Mot Behav. 1989;21:225–244.



41 Bisdorff AR, Wolsley CJ, AnastasopoulosD, et al. The perception of body verticality(subject postural vertical) in peripheraland central vestibular disorders. Brain.1996;119:1523–1534.

42 Duncan PW, Weiner DK, Chandler J, Stu-denski S. Functional reach: a new clinicalmeasure of balance. J Gerontol. 1990;45:M192–M197.

43 Newton RA. Validity of the Multi-directional Reach Test: a practical mea-sure for limits of stability in older adults.J Gerantol A Biol Sci Med Sci.2001; 56:M248–M252.

44 Jonsson E, Henriksson M, Hirschfeld H.Does the functional reach test reflect sta-bility limits in elderly people? J RehabilMed. 2002;35:26–30.

45 Crenna P, Frigo C. A motor programme forthe initiation of forward-oriented move-ments in humans. J Physiol. 1991;437:635–653.

46 Rogers MW, Hedman LD, Pai Y-C. Kineticanalysis of dynamic transitions in stancesupport accompanying voluntary leg flex-ion movements in hemiparetic adults.Arch Phys Med Rehabil. 1993;74:19–25.

47 Horak FB, Anderson M. Preparatory pos-tural activity associated with movement[abstract]. Phys Ther. 1980;60:580.

48 Massion J, Woollacott MH. Posture andequilibrium. In: Bronstein AM, Brandt T,Woollacott MH, eds. Clinical Disordersof Balance, Posture and Gait. London,United Kingdom: Arnold; 1996:1–18.

49 Horak FB, Jacobs JV, Tran VK, Nutt JG.The push and release test: an improvedclinical postural stability test for patientswith Parkinson’s disease [abstract]. MovDisord. 2004;19(Suppl 9):S170.

50 Maki BE, McIlroy WE. The role of limbmovements in maintaining upright stance:The “change-in-support” strategy. PhysTher. 1997;77:488–507.

51 Shumway-Cook A, Horak FB. Assessing theinfluence of sensory interaction on bal-ance: suggestion from the field. Phys Ther.1986;66:1548–1550.

52 Huxham FE, Goldie PA, Patla AE. Theoret-ical considerations in balance assessment.Aust J Physiother. 2001;47:89–100.

53 Paquette C, Paquet N, Fung J. Aging affectscoordination of rapid head motions withtrunk and pelvis movements during stand-ing and walking. Gait Posture. 2006;24:62–69.

54 Shumway-Cook A, Woollacott MH. MotorControl: Theory and Practical Applica-tions. Baltimore, MD: Williams & Wilkins;2001.

55 Mathias S, Nayak USL, Isaacs B. Balancein elderly patients: the “get-up and go”test. Arch Phys Med Rehabil. 1986;67:387–389.

56 Shumway-Cook A, Horak FB. Vestibular re-habilitation: an exercise approach to man-aging symptoms of vestibular dysfunction.Semin Hearing. 1989;10:196–208.

57 Shumway-Cook A, Horak FB. Rehabilita-tion strategies for patients with vestibulardeficits. Neurol Clin. 1990;8:441–457.

58 Powell LE, Myers AM. The Activities-specific Balance Confidence (ABC) Scale.J Gerontol A Biol Sci Med Sci. 1995;50:M28–M34.

59 Myers AM, Fletcher PC, Myers AH, SherkW. Discriminative and evaluative proper-ties of the Activities-specific Balance Con-fidence (ABC) Scale. J Gerontol. 1998;53:M287–M294.

60 Shrout PE, Fleiss JL. Interclass correlation:uses in assessing rater reliability. PsycholBull. 1979;86:420–428.

61 Portney LG, Watkins MP. Foundations ofClinical Research: Applications to Prac-tice. Upper Saddle River, NJ: Prentice HallHealth; 2000.

62 Berg KO, Wood-Dauphinee SL, Williams JI,Maki B. Measuring balance in the elderly:validation of an instrument. Can J PublicHealth. 1992;83(Suppl 2):S7–S11.

63 Tinetti ME, Mendes de Leon CF, DoucetteJT, Baker DI. Fear of falling and fall-relatedefficacy in relationship to functioningamong community-living elders. J Geron-tol. 1994;49:140–147.

64 Close JC, Lord SL, Menz HB, SherringtonC. What is the role of falls? Best Pract ResClin Rheumatol. 2005;19:913–935.

65 Lord SR, Menz HB, Tiedemann A. A phys-iological profile approach to falls riskassessment and prevention. Phys Ther.2003;83:237–252.

66 Woollacott MH, Shumway-Cook A. MotorControl: Translating Research Into Clini-cal Practice. Philadelphia, PA: LippincottWilliams & Wilkins; 2007.

67 Runge CF, Shupert CL, Horak FB, Zajac F.Role of vestibular information in initiationof rapid postural responses. Exp BrainRes. 1998;122:403–412.

68 Shupert CL, Horak FB. Effects of vestibularloss on head stabilization in response tohead and body perturbations. J Vestib Res.1996;6:423–437.

69 Jobges M, Heuschkel G, Pretzel C, et al.Repetitive training of compensatory steps:a therapeutic approach for postural insta-bility in Parkinson’s disease. J Neurol Neu-rosurg Psychiatry. 2003;75:1682–1687.

70 Shumway-Cook A, Horak FB, Yardley L,Bronstein AM. Rehabilitation of balancedisorders in the patient with vestibular pa-thology. In: Bronstein AM, Brandt T, Wool-lacott MH, eds. Clinical Disorders of Bal-ance, Posture, and Gait. London, UnitedKingdom: Arnold, Div of Hodder HeadlinePLC; 1996:213–220.

71 Berg KO, Wood-Dauphinee S, Williams JI.The Balance Scale: reliability assessmentwith elderly residents and patients with anacute stroke. Scand J Rehabil Med. 1995;27:27–36.

72 Cipriany-Dacko LM, Innerst D, et al. In-terrater reliability of the Tinetti BalanceScores in novice and experienced physicaltherapy clinicians. Arch Phys Med Reha-bil. 1997;78:1160–1164.

73 Anacker SL, Di Fabio RP. Influence ofsensory inputs on standing balance incommunity-dwelling elders with a recenthistory of falling. Phys Ther. 1992;72:575–581; discussion 581–574.

74 Wrisley DM, Walker ML, Echternach JL,Strasnick B. Reliability of the DynamicGait Index in people with vestibular dis-orders. Arch Phys Med Rehabil. 2003;84:1528–1533.

75 Folio R, Fewell RP. Peabody Developmen-tal Motor Scales, Second Edition (PDMS-2). Los Angeles, CA: Western Psychologi-cal Services; 2000.

76 Wang HH, Liao HF, Hsieh CL. Reliability,sensitivity to change, and responsivenessof the Peabody Developmental MotorScales—Second Edition for children withcerebral palsy. Phys Ther. 2006;86:1351–1359.



Documents

The Balance Evaluation Systems Test