Reliability and validity of the Balance Evaluation Systems Test (BESTest) in people with subacute stroke

doi: 10.2522/ptj.20130558Originally published online June 12, 2014

Published online June 12, 2014PHYS THER. Rumpa BoonsinsukhSaengsirisuwan, Nitaya Viriyatharakij, Fay B. Horak and Butsara Chinsongkram, Nithinun Chaikeeree, VitoonStrokeSystems Test (BESTest) in People With Subacute Reliability and Validity of the Balance Evaluation

http://ptjournal.apta.org/content/early/2014/09/10/ptj.20130558found online at: The online version of this article, along with updated information and services, can be

Collections

Tests and Measurements     Stroke (Neurology)    

Stroke (Geriatrics)     Balance    

in the following collection(s): This article, along with others on similar topics, appears

E-mail alerts to receive free e-mail alerts hereSign up

corrections and replace the original author manuscript. : edited and typeset versions of articles that incorporate any authorPage proofs

  readers almost immediate access to accepted papers.

PTJaccepted for publication but have not yet been copyedited or typeset. This allows : PDF versions of manuscripts that have been peer-reviewed andAuthor manuscripts

  publishes 2 types of Online First articles: PTJ). PTJ (Physical Therapy

Online First articles are published online before they appear in a regular issue of

by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

Reliability and Validity of the BalanceEvaluation Systems Test (BESTest) inPeople With Subacute StrokeButsara Chinsongkram, Nithinun Chaikeeree, Vitoon Saengsirisuwan,Nitaya Viriyatharakij, Fay B. Horak, Rumpa Boonsinsukh

Background. The Balance Evaluation Systems Test (BESTest) is a new clinicalbalance assessment tool, but it has never been validated in patients with subacutestroke.

Objective. The purpose of this study was to examine the reliability and validity ofthe BESTest in patients with subacute stroke.

Design. This was an observational reliability and validity study.

Methods. Twelve patients participated in the interrater and intrarater reliabilitystudy. Convergent validity was investigated in 70 patients using the Berg BalanceScale (BBS), Postural Assessment Scale for Stroke (PASS), Community Balance andMobility Scale (CB&M), and Mini-BESTest. The receiver operating characteristic curvewas used to calculate the sensitivity, specificity, and accuracy of the BESTest, Mini-BESTest, and BBS in classifying participants into low functional ability (LFA) and highfunctional ability (HFA) groups based on Fugl-Meyer Assessment motor subscalescores.

Results. The BESTest showed excellent intrarater reliability and interrater reliabil-ity (intraclass correlation coefficient�.99) and was highly correlated with the BBS(Spearman r�.96), PASS (r�.96), CB&M (r�.91), and Mini-BESTest (r�.96), indicat-ing excellent convergent validity. No floor or ceiling effects were observed with theBESTest. In contrast, the Mini-BESTest and CB&M had a floor effect in the LFA group,and the BBS and PASS demonstrated responsive ceiling effects in the HFA group. Inaddition, the BESTest showed high accuracy as the BBS and Mini-BESTest in separat-ing participants into HFA and LFA groups.

Limitation. Whether the results are generalizable to patients with chronic strokeis unknown.

Conclusion. The BESTest is reliable, valid, sensitive, and specific in assessingbalance in people with subacute stroke across all levels of functional disability.

B. Chinsongkram, MS, Division ofPhysical Therapy, Faculty ofHealth Science, SrinakharinwirotUniversity, Nakhon Nayok, Thai-land, and Faculty of Physical Ther-apy, Rangsit University, PathumThani, Thailand.

N. Chaikeeree, MS, Division ofPhysical Therapy, Faculty ofHealth Science, SrinakharinwirotUniversity, and Department ofPhysiology, Faculty of Science,Mahidol University, Bangkok,Thailand.

V. Saengsirisuwan, PhD, Depart-ment of Physiology, Faculty of Sci-ence, Mahidol University.

N. Viriyatharakij, PhD, Division ofPhysical Therapy, Faculty ofHealth Science, SrinakharinwirotUniversity.

F.B. Horak, PhD, Balance Disor-ders Laboratory, Neurological Sci-ences Institute, Oregon Healthand Science University, Beaverton,Oregon.

R. Boonsinsukh, PhD, Division ofPhysical Therapy, Faculty ofHealth Science, SrinakharinwirotUniversity, 63 moo 7 Rungsit-Nakhonnayok, Ongkharuk, Nak-hon Nayok, 26120. Address allcorrespondence to Dr Boonsin-sukh at: [email protected].

[Chinsongkram B, Chaikeeree N,Saengsirisuwan V, et al. Reliabilityand validity of the Balance Evalu-ation Systems Test (BESTest) inpeople with subacute stroke. PhysTher. 2014;94:xxx–xxx.]

© 2014 American Physical TherapyAssociation

Published Ahead of Print:June 12, 2014

Accepted: June 2, 2014Submitted: November 21, 2013

Research Report

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

November 2014 Volume 94 Number 11 Physical Therapy f 1 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

Balance control is a complexintegration of multiple systems,including sensory and motor

systems, sensorimotor integration,and high-level premotor process-ing—all of which can be affected fol-lowing stroke.1 Laboratory studieshave shown that muscle paresis andspasticity can impair postural align-ment, base of foot support, andforces necessary to correct posturalequilibrium.2–4 Most patients withstroke also demonstrate decreasedankle proprioception and somato-sensation at the feet, which are crit-ical for balance control.2,4 In addi-tion, patients with stroke havedifficulties in central processing forsensory integration and sensoryreweighting.5 They demonstratemore postural sway in the conditionsof altered somatosensory informa-tion and altered vision,6 such aswalking on uneven surfaces andwalking in the dark. Anticipatorypostural adjustments and automaticpostural responses to restore thebody’s equilibrium during perturba-tions are reduced on the pareticside.7–10 Cognitive impairments canlead to inadequate allocation ofattention and to greater fall risk inthis population.2,11 These studiessuggest that postural control impair-ments in patients with stroke arisefrom diverse mechanisms involvedin balance control. Therefore, a bal-ance assessment that could captureall underlying balance impairmentsis necessary.

At present, balance evaluation forassessing patients with stroke in clin-ical settings involves ordinal scales toassess a set of functional tasks thatrequire balance while maintainingposture or changing position.12

Examples of the most commonlyused functional scales in patientswith stroke include the Berg BalanceScale (BBS) and the Postural Assess-ment Scale for Stroke Patients(PASS).13 These functional balancescales are useful in reporting the

presence of balance impairments,but they cannot identify the type ofimpairments underlying balanceproblems.

Recently, the Balance Evaluation Sys-tems Test (BESTest) has been devel-oped based on a conceptual modelof balance control in which 6 differ-ent systems or domains contribute tobalance control. The BESTest evalu-ates the following systems: biome-chanical constraints, stability limitsand verticality, anticipatory posturaladjustments, automatic posturalresponses, sensory organization, andstability in gait.14 The 36 items of theBESTest are scored from 0 (severeimpairment) to 3 (no impairment).The total score is calculated as a per-centage of the points scored out of108 total points.14 The advantage ofthe BESTest is that it includes tasksinvolving many different mecha-nisms underlying the control of bal-ance so that clinicians can determinethe type of balance retraining inter-vention specific to the underlyingcauses of balance problems. TheBESTest has shown excellent interra-ter reliability and validity and thepotential to differentiate types of bal-ance deficits in patients with Parkin-son disease (PD), vestibular loss, andperipheral neuropathy.14 However,the BESTest has not yet been vali-dated to assess balance performancein patients with stroke.

The Mini-BESTest, the shorter ver-sion of the BESTest, was developedto reduce the assessment time.15

Item redundancy and the first 2 sec-tions (biomechanical constraints andstability limits with verticality) fromthe BESTest were removed based onRasch analysis to yield the assess-ment of a common construct,“dynamic balance,” in the Mini-BESTest.15 The 14-item Mini-BESTestis scored on a 3-level ordinal scalefrom 0 (poor balance performance)to 2 (no balance impairment). Sev-eral studies have shown that the

Mini-BESTest also was reliable andvalid16–18 and useful for fall predic-tion in patients with PD.19,20 The useof the Mini-BESTest in community-dwelling people with chronic strokehas recently been reported, withexcellent interrater and intraraterreliability and high correlations withother balance measures such as theBBS and the Timed “Up & Go” Testbut without the floor and ceilingeffects seen in these traditional bal-ance tests.21 However, the feasibilityof the Mini-BESTest for assessing bal-ance across lower levels of subacutestroke impairment severity has notbeen reported.

The purposes of this study were:(1) to assess the reliability and con-vergent validity of the BESTest inpatients with subacute stroke and(2) to determine whether theBESTest could be used to identifypatients with low and high func-tional ability, as classified with theFugl-Meyer Stroke Assessment motorsubscale (FM-motor). We hypothe-sized that the BESTest would bereliable and well correlated withother balance scales such as the BBS,PASS, and Mini-BESTest. Further-more, we predicted that the BESTestwould have fewer floor and ceilingeffects than the other clinical bal-ance tests, so that the BESTest willbe more feasible for people withstroke across a wider range ofmotor disability. Finally, we hypoth-esized that the lack of musculoskel-etal and limits of stability/verticalitydomains in the Mini-BESTest maylead to a floor effect in patientswho have low functional abilitybecause these domains are com-monly affected in patients with sub-acute stroke.

MethodParticipantsParticipants were patients with sub-acute stroke (defined as stroke ofless than 4 months since onset22)referred for physical therapist ser-

Validation of BESTest in Stroke

2 f Physical Therapy Volume 94 Number 11 November 2014 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

vices at Prasart Neurological Insti-tute in Thailand from November 1,2012, to August 31, 2013. Individualswere included in this study if theypassed the screening tests by a phy-sician using the following criteria:diagnosis of cerebral hemorrhage orcerebral infarction with stable medi-cal conditions, age between 25 and90 years, first unilateral hemisphericstroke, and able to follow instruc-tions to complete the assessment.Individuals were excluded from thestudy if they had cognitive impair-ment (defined as having a Mini-Mental State Examination [MMSE]score of less than 24),23,24 cerebralaneurysm, a lesion at the brain steminvolving the sleep-wake and respi-ratory control center or cerebellum,aphasia as diagnosed by a physicianusing a Thai adaptation of the West-ern Aphasia Battery,25 a neurologicaldisorder other than stroke, or thepresence of a major peripheral neu-ropathy or musculoskeletal problemthat was sufficient to disturb bal-ance. All participants signed a con-sent form approved by the HumanResearch Protection Committee atPrasart Neurological InstituteResearch Center, Thailand.

Outcome MeasuresTo determine whether the balancetools could be used to separate par-ticipants into high- and low-functioning groups, the FM-motorwas selected to measure functionalrecovery from stroke.26 TheFM-motor is scored from 0 to 100points, with each item scored on a3-point ordinal scale ranging from 0(cannot perform) to 2 (performsfully). The FM-motor score is used toclassify patients with stroke into 4groups according a level of disability:a score of 0 to 35 represents severedisability, a score of 36 to 55 repre-sents moderately severe disability, ascore of 56 to 79 represents moder-ate disability, and a score of 80 orhigher represents mild disability.27

In this study, we referred to patients

with stroke who had FM-motorscores between 0 and 55 as the lowfunctional ability group (LFA group)and to those who had FM-motorscores greater than 55 as the highfunctional ability group (HFAgroup).

Five balance evaluation scales—BESTest, Mini-BESTest, BBS, PASS,and Community Balance and Mobil-ity Scale (CB&M)—were adminis-tered to each participant. The BBS isa widely used clinical scale for func-tional balance in the clinic and iscurrently considered as a referencestandard for assessing balance inpatients with stroke. The BBS con-sists of 14 functional balance itemsthat focus on the ability to maintain aposition and postural adjustments tovoluntary movements.28 The totalpossible score on the BBS is 56points, with ratings ranging from 0(unable to perform) to 4 (able tocomplete the task).13 The PASS is a12-item test designed to measurebalance function in patients withstroke, especially those with verypoor motor function. The 4-pointitems, creating the total score of36, assess functional balance per-formance in maintaining or chang-ing lying, sitting, and standing pos-tures.29 The CB&M consists of 19tasks, including advanced func-tional balance and mobility activitiesaimed at ambulatory or community-dwelling people with stroke.30 Itemsare scored on ordinal scale rangingfrom 0 (unable to perform) to 5 (ableto perform independently).

ProcedureReliability. Interrater and intra-rater reliability were determined by5 physical therapists. Raters were aconvenient sample of 3 physicaltherapists from the stroke rehabilita-tion center with stroke rehabilitationexperience of 1, 5, and 10 years, and2 PhD physical therapist studentsfrom Srinakharinwirot Universityand Mahidol University. Prior to the

study, all raters participated in 2training workshops on using theBESTest. The training workshopswere carried out by a therapist whoreceived training from the BESTestdeveloper. Each item of the BESTestwas demonstrated on individualswho were healthy, and the ratingcriteria were discussed. All ratersalso scored each participant’s perfor-mance from a demonstration video,and rating criteria were discussedagain if there was a discrepancy.When there were differences in therating scores, the raters were askedto watch the videotape with thetrainer, who encouraged the discus-sion until all raters reachedagreement.

After the training workshops, aBESTest reliability trial was per-formed in 12 patients with stroke.The sample size calculation for thisstudy was based on a power of 0.80and an alpha level of .05. A null intra-class correlation coefficient (ICC) of.6031 and expected correlation coef-ficients in our study were based onthe correlation coefficient of r�.93found in a previous balance measurereliability study.32 The BESTest wasadministered by one rater (B.C.), andthe procedures were videorecorded. The evaluation was per-formed in a laboratory setting atPrasart Neurological Institute, andvideotape recording was performedin the same view for all participants.The rater used the same testinginstructions to standardize perfor-mance. Before testing, the vital signsof the participants were evaluated toensure stable medical status. All par-ticipants received the same verbalinstructions. Participants were per-mitted to rest as needed during thetest.

All raters scored each participant’sperformance from videotape on 2separate occasions. The second scor-ing was performed 7 days after thefirst scoring. Intrarater reliability of

Validation of BESTest in Stroke

November 2014 Volume 94 Number 11 Physical Therapy f 3 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

total scores and section scores wasassessed by comparing the score ofoccasion 1 and score of occasion 2for each rater. Interrater reliabilitywas determined by comparing thescores from occasion 1 for all 5 rat-ers. Each rater scored the videotapeon separate scoring sheets for eachoccasion and did not discuss scoringamong participants and occasions.

Validity. Prior to the validitystudy, the raters (B.C., N.C.)received additional training for usingthe Mini-BESTest, BBS, PASS, CB&M,and FM-motor. The convergent valid-ity sample size was calculated from anull correlation coefficient of .5033

and acceptable correlation of .80from a previous study,32 yielding atleast 29 participants per group. Aftersigning the approved consent forms,each participant’s demographic andclinical information, including Bar-thel Index (BI) score and FM-motorscore, was recorded by one rater(N.C.). Seventy patients, 35 in eachgroup, participated in this study. Abalance assessment of each partici-pant was then performed by the sec-ond rater (B.C.), who was blinded tothe basic characteristics and theFM-motor scores of the participants.The sequence of the balance testswas randomly performed followingthe order of the starting position:from lying to sitting to standing andwalking. Any test item that wasduplicated between the tests wasperformed only once and scoredusing the criteria for each test.34 Theevaluation was performed in thesame laboratory setting, and all par-ticipants received the same verbalinstructions. Rest was allowed dur-ing performance of each item toavoid fatigue. Total assessment timewas approximately 1.5 hours, but ifthe testing could not be completedin 1 day, it was continued on thenext day. To ensure the accuracy ofthe scoring, the rater (B.C.) reviewedand scored each participant’s testperformance from the videotape

again. In case of disagreementbetween concurrent rating andvideo rating, the second rater (N.C.)was asked to score those items fromthe videotape to decide the finalscore.

Data AnalysisA descriptive statistical analysis ofdemographic and baseline clinicalcharacteristics of the participantswas conducted. An independent-sample t test was used to compareage, time since stroke, BI score, andBBS score, whereas the Mann-Whitney U test was used to compareMMSE score, FM-motor score, andother balance measurement scoresbetween the LFA and HFA groups.Interrater and intrarater reliabilityvalues were calculated based on theparticipants’ ICCs.35 The ICC model2,k was used for interrater reliability,and the ICC model 3,k was used forintrarater reliability. An ICC value of.80 or higher indicates excellent cor-relation (good reliability), values of.60 to .80 indicate adequate correla-tion (moderate reliability), and val-ues of .40 to .60 indicate poor cor-relation (weak reliability).31,36

To determine the convergent validityof the BESTest, the strength of therelationship among the BESTest,Mini-BESTest, PASS, BBS, and CB&Mwas examined using Spearman rankorder correlations. Correlation coef-ficients of .00 to .49 were inter-preted as poor, those of .50 to .79 asmoderate, and those of 0.8 or higheras excellent.33 The floor and ceilingeffects were calculated as the per-centage of the sample scoring theminimum or maximum possiblescores, respectively. Ceiling andfloor effects of 20% or greater areconsidered significant.37 In thisstudy, we also recorded the “respon-sive ceiling effect,” which representsthe 20% or greater percentage of par-ticipants who scored within the top10% of the test and allows a suffi-cient number of scores to measure

change over time.34 The percentageof minimum and maximum scoresfor each item of the BESTest wascalculated as the percentage of par-ticipants who received a minimumscore (0) or a maximum score (3) foreach item. These percentages repre-sented the difficulty of the item, suchthat a high percentage (�80%) ofminimum score reflected a difficultitem and a high percentage of maxi-mum score indicated an easy item.

Receiver operating characteristic(ROC) curves were used to deter-mine the relative performance of theBESTest score, the Mini-BESTestscore, and the BBS score for classify-ing participants into 2 groups basedon their FM-motor score to discrimi-nate between low functional ability(FM-motor score �55) and highfunctional ability (FM-motor score�55). The accuracy for each balancetest was assessed using the areaunder the curve (AUC), which canbe interpreted as the probability ofcorrectly identifying participantswith low or high function. An AUCvalue of 0.9 and above indicates highaccuracy, 0.7 to 0.9 indicates moder-ate accuracy, 0.5 to 0.7 indicates lowaccuracy, and 0.5 and below indi-cates test due to chance.38 The cut-off point value was chosen by select-ing the score that provides the bestbalance between high sensitivity andhigh specificity.35 Furthermore, like-lihood ratios and accuracy were usedto determine posttest probabilitiesthat helped to improve certainty forconfirming the diagnosis or aban-doning it. A positive likelihood ratio(LR�) was the probability of a par-ticipant having a score above thecutoff point, whereas a negativelikelihood ratio (LR�) was the prob-ability of a participant having ascore below the cutoff point. AnLR� over 5 and an LR� less than 0.2indicate that the test is useful due toits high probability of correctly clas-sifying patients into functional abilitygroups, but likelihood ratios close to

Validation of BESTest in Stroke

4 f Physical Therapy Volume 94 Number 11 November 2014 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

1.00 indicate that the test are useless,as the probability of correctly orincorrectly classifying patients intogroups is the same.35 The accuracyof the selected cutoff point in clas-sifying patients into functional abil-ity groups was calculated as theproportion of the participants inthe correctly identified group whenusing that cutoff point for allparticipants.35

Role of the Funding SourceThis study was supported by theThailand Research Fund, the Officeof Higher Education Commission,and Srinakharinwirot University(grant no. RSA5580002).

ResultsReliabilityTwelve patients with subacute hem-orrhagic or ischemic stroke (4female, 8 male; average age�58.24years) participated in the reliabilitytesting. The intrarater reliability ofthe BESTest total scores was excel-lent, with an ICC of .99 (95% confi-dence interval [95% CI]�.99, 1.00)

and subsection ICCs ranging from.95 (95% CI�.90, .98) to .99 (95%CI�.98, .99). The interrater reliabil-ity of the BESTest total scores alsowas excellent, with an ICC of .99(95% CI�.98, .99) and subsectionICCs ranging from .87 (95% CI�.72,.96) to .98 (95% CI�.95, .99).

Convergent ValidityThe eFigure (available at ptjournal.apta.org) depicts the number of eli-gible participants, included partici-pants, and excluded participantswith the reasons for exclusion. Onehundred thirteen patients with sub-acute stroke met the eligibility crite-ria, and 70 patients consented to par-ticipate in the validity testing. Of the43 patients who were not includedin this study, 30 did not require con-tinued rehabilitation, 5 had cognitiveimpairment or severe motor aphasia,2 had severe internal carotid arteryocclusion that required urgent sur-gery, and 6 had major musculoskele-tal problems, including osteoarthritisand gouty arthritis. The majority ofthe participants (n�67) completed

the validity testing within 1 day.Demographic and clinical character-istics of participants in the validitystudy are shown in Table 1. Partici-pants in the LFA and HFA groupswere similar in age, sex, and cogni-tive impairment as measured withthe MMSE, but those in the LFAgroup had a longer time since strokeand significantly lower scores on allof the clinical balance tests.

The convergent validity of the BEST-est in patients with subacute strokewas assessed with clinical referencescales that have been validated inpopulations with stroke (ie, BBS,PASS, CB&M, and Mini-BESTest). Fig-ure 1 shows high correlations of thetotal scores from the BESTest withthe BBS (r�.96), PASS (r�.96),CB&M (r�.91), and Mini-BESTest(r�.96), indicating excellent conver-gent validity.

Score Distribution Across Levelsof Functional AbilityThe distribution of participants’scores for the balance measures

Table 1.Demographic and Clinical Characteristics of Participants in the Validity Study, Classified by Functional Abilitya

Characteristic All Participants (N�70)

Functional Ability Group

Low (n�35) High (n�35)

Age (y) 57.01 (12.23) 54.85 (12.09) 59.17 (12.16)

Sex (male/female), n 38/32 21/14 17/18

Time since stroke (mo) 1.11 (2.00) 1.74 (2.60)* 0.48 (0.75)*

Type of stroke (ischemic/hemorrhagic), n 54/16 23/12 31/4

Affected side (right/left), n 39/31 22/13 17/18

MMSE (/30) 28.00 (2.39) 27.06 (2.38) 28.77 (2.10)

BI (/100) 61.93 (24.76) 48.86 (18.87)† 75.00 (23.17)†

FM-motor (/100) 55.53 (34.00) 24.37 (12.59)† 86.69 (13.72)†

BESTest (/100%) 41.70 (28.19) 23.89 (18.44)† 59.52 (24.82)†

Mini-BESTest (/28) 9.07 (8.53) 4.06 (5.44)† 14.09 (8.14)†

BBS (/56) 29.46 (19.17) 18.37 (15.92)* 40.57 (15.49)*

PASS (/36) 26.70 (7.79) 22.03 (7.7)* 31.37 (4.33)*

CB&M (/96) 11.64 (18.98) 1.37 (3.95)† 21.91 (22.32)†

a All values are presented as mean (SD), unless otherwise noted. Significant difference between low and high functional ability groups (* P�.01, † P�.001).Functional ability was classified with the Fugl-Meyer Stroke Assessment–motor subscale (FM-motor) as low (0–55) or high (�55). MMSE�Mini-Mental StateExamination, BI�Barthel Index, BESTest�Balance Evaluation Systems Test, Mini-BESTest�Mini-Balance Evaluation Systems Test, BBS�Berg Balance Scale,PASS�Postural Assessment Scale for Stroke Patients, CB&M�Community Balance and Mobility Scale.

Validation of BESTest in Stroke

November 2014 Volume 94 Number 11 Physical Therapy f 5 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

across levels of functional ability isshown in Figure 2. The distributionof scores from all observed scalesexcept the CB&M covered the wholerange of functional ability. However,for the Mini-BESTest and PASS,scores were clustered at one end ofthe graph, indicating the potentialfloor or ceiling effects. The analysisof the floor effects demonstrated that34.3% (n�12) of the participants inthe LFA group (17% of all partici-pants with subacute stroke) received

the lowest possible score of 0 on theMini-BESTest, suggesting the flooreffect of the Mini-BESTest specifi-cally in the LFA group. Likewise,77.1% (n�27) of the participants inthe LFA group and 11.4% (n�4) ofthe participants in the HFA group(44.3% of all participants with sub-acute stroke) had a score of 0 on theCB&M. None of the participants hada score of 0 on the BESTest, BBS, orPASS.

For the ceiling effect, 4.3% (n�3) ofthe participants received the maxi-mum score on the BBS and PASS.However, when calculating by usingthe scores within the top 10%, 25.7%of all participants (5.7% of the par-ticipants in the LFA group and 45.7%of the participants in the HFA group)scored within the top 10% on theBBS, indicating a responsive ceilingeffect especially in the HFA group.Similarly, 37.1% of all participants(14.3% of the participants in the LFAgroup and 60.0% of the participantsin the HFA group) received a scorewithin the top 10% on the PASS. Tothe contrary, no responsive ceilingeffect was observed in the rest of thebalance scales, as very few partici-pants received a score within the top10% (BESTest: 4.3%, Mini-BESTest:4.3%, and CB&M: 1.4%).

Item Scores of the BESTestThe scores of each item and sectionof the BESTest compared betweenthe LFA and HFA groups are shownin Table 2. In general, the sectionscores were significantly lower inthe LFA group, suggesting that theBESTest was able to differentiatebetween the 2 groups. Several itemsin the BESTest were too hard to per-form for the LFA group, as shown bythe percentage of minimum score ofmore than 80%. These itemsincluded sit on floor, rise to toe,stand on one leg, alternate stairtouching, compensatory posturalresponse in all directions, stand onfoam with eyes closed, and all itemsin the “stability in gait” section, withthe exception of change in gaitspeed. In contrast, there were only 2items (base of support and sittingverticality) that were shown to beeasy (�80% of maximum score) inboth LFA and HFA groups.

Comparison of the percentage ofminimum and maximum scoresbetween the LFA and HFA groupsalso suggested the feasibility of eachitem to be used in all levels of func-

Figure 1.Relationships between the total scores of the Balance Evaluation System Test (BESTest)and scores from other clinical balance scales: Mini-Balance Evaluation System Test(Mini-BESTest), Berg Balance Scale (BBS), Postural Assessment Scale for Stroke Patients(PASS), and Community Balance and Mobility Scale (CB&M). High concurrent validityof the BESTest with other clinical scales of balance was found, as well as ceiling effectsfor the BBS and PASS and a floor effect for the CB&M.

Validation of BESTest in Stroke

6 f Physical Therapy Volume 94 Number 11 November 2014 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

tional ability. For example, 31.4% ofthe participants in the LFA groupcompared with 71.4% of the partici-pants in the HFA group received themaximum score on item 6 (laterallean). A similar trend was evidentwith the minimum score, wherethere was a decrease in percentageof minimum score in the HFA groupcompared with the LFA group. Sucha pattern was found for all items ofthe BESTest, except for 3 items (baseof support and sitting verticality, leftand right).

Identification of FunctionalAbility LevelThe ROC plots for the BESTest, Mini-BESTest, and BBS are shown in Fig-ure 3. All 3 scales showed high AUCsin identifying patients with stroke inthe LFA and HFA groups: 0.88 for theBESTest, 0.85 for the Mini-BESTest,and 0.85 for the BBS (Tab. 3). Thesuggested cutoff scores for low func-tional ability were 49% for theBESTest, 9/28 for the Mini-BESTest,and 19/56 for the BBS (Tab. 3). Basedon these cutoff scores, comparablesensitivity and specificity was foundin all 3 measures, with specificityhighest for the BESTest, whereas theBBS had the highest sensitivity butthe lowest specificity. All 3 scalesalso showed high posttest accuracy(0.77–0.81) when applying the sug-gested cutoff point in classifying par-ticipants into the LFA and HFAgroups. Although all 3 tests showedsimilar AUC, sensitivity, specificity,and posttest accuracy, the BESTestand Mini-BESTest had LR� ratiosover 5 and LR� ratios close to 0.2,suggesting the BESTest and Mini-BESTest had higher probability ofcorrectly classifying patients accord-ing to level of functional ability com-pared with the BBS.

DiscussionThis is the first study to examinewhether the BESTest is reliable andvalid to assess balance impairmentsin patients with subacute stroke. Our

results confirmed that the BESTesthas excellent interrater and intra-rater reliability, as well as excellentconvergent validity, compared with

commonly used clinical balancescales in patients with stroke. Thehigh intrarater and interrater reliabil-ity of the BESTest in patients with

Figure 2.Distribution of all participants’ scores from the Balance Evaluation System Test(BESTest), Mini-Balance Evaluation System Test (Mini-BESTest), Berg Balance Scale(BBS), Postural Assessment Scale for Stroke Patients (PASS), and Community Balanceand Mobility Scale (CB&M) across levels of functional ability as determined with theFugl-Meyer Stroke Assessment–motor subscale (FM-motor). The figure shows that theFM-motor cannot predict balance function, except at lowest and highest levels. Ceilingand floor effects were found for last 2 tests.

Validation of BESTest in Stroke

November 2014 Volume 94 Number 11 Physical Therapy f 7 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

Table 2.Item Scores on the Balance Evaluation Systems Test (BESTest)a

BESTest

LFA Group HFA Group

Median

%N

Median

%N

Section Item Minimum Maximum Minimum Maximum

I Biomechanical constraints 31.43 (17.08)* 57.52 (22.00)*

1. Base of support 3 0.0 82.9 3 0.0 91.4

2. CoM alignment 1 37.1 5.7 2 2.9 48.6

3. Ankle strength/ROM 0 71.4 2.9 1.5 28.6 8.6

4. Hip/trunk strength 0 80.0 0.0 0 60.0 8.6

5. Sit on floor/stand up 0 85.7 5.7 2 42.9 25.7

II Limits of stability 50.95 (19.57)* 74.29 (14.93)*

6A. Sitting verticality, right† 3 5.7 82.9 3 0.0 100.0

6B. Lateral lean, right† 2 5.7 31.4 3 0.0 71.4

7. FR forward 0 65.7 0.0 2 17.1 5.7

8. FR right† 0 68.6 2.9 2 22.9 8.6

III Anticipatory adjustments 18.41 (20.73)* 57.46 (26.16)*

9. Sit to stand 2 28.6 40.0 3 0.0 85.7

10. Rise to toe 0 85.7 0.0 2 25.7 22.9

11. Stand on one leg, right† 0 88.6 0.0 1 40.0 11.4

12. Alternate stair touching 0 91.4 0.0 2 31.4 31.4

13. Standing arm raise 0 68.6 20.0 3 17.1 57.1

IV Postural responses 12.86 (26.40)* 60.98 (36.7)*

14. In-place, front 0 71.4 20.0 3 14.3 71.4

15. In-place, back 0 74.3 14.3 3 17.1 62.9

16. Compensatory, front 0 91.4 5.7 2 31.4 42.9

17. Compensatory, back 0 91.4 5.7 2 37.1 37.1

18. Compensatory, right† 0 91.4 5.7 2 42.9 31.4

V Sensory orientation 23.24 (28.58)* 62.29 (29.90)*

19A. Firm surface, EO 1 45.7 34.3 3 2.9 80.0

19B. Firm surface, EC 0 54.3 14.3 3 2.9 60.0

19C. Foam surface, EO 0 77.1 2.9 2 17.1 34.3

19D. Foam surface, EC 0 91.4 0.0 0.5 51.4 8.6

20. Incline, EC 0 62.9 17.1 3 20.0 60.0

VI Stability in gait 6.26 (15.61)* 44.35 (31.68)*

21. Gait, level surface 0 82.9 0.0 2 25.7 5.7

22. Change in gait speed 0 80.0 0.0 2 25.7 22.9

23. Walk head turn, lateral 0 82.9 0.0 1 42.9 14.3

24. Walk with pivot turns 0 91.4 2.9 2 28.6 22.9

25. Step over obstacles 0 91.4 0.0 2 34.3 25.7

26. TUG 0 91.4 0.0 2 31.4 2.9

27. TUG with dual task 0 82.9 0.0 1 25.7 0.0

a All section scores are presented as mean (SD). LFA�low functional ability, HFA�high functional ability, FR�functional reach, EO�eyes open, EC�eyesclosed, TUG�Timed “Get Up and Go” test, CoM�center of mass, ROM�range of motion. * Significant difference between LFA and HFA groups at P�.01.† Item that evaluated both right and left sides, but with similar results on each side, so data for only the right side are presented.

Validation of BESTest in Stroke

8 f Physical Therapy Volume 94 Number 11 November 2014 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

subacute stroke agreed with previ-ous studies in patients with neuro-logical conditions or healthy elderlypeople.14,34 For example, the ICC fortotal score interrater reliability was.91, with section score ICCs rangingfrom .79 to .96 in patients with neu-rological diseases and healthy peo-ple.14 Similarly, the ICC for totalscore interrater reliability was .96,with section score ICCs rangingfrom .79 to .96 in patients with PD.39

The high correlations (r�.90)between the BESTest and other clin-ical balance measures indicated that,to some extent, the BESTest mea-sures similar balance constructs asthe BBS, PASS, CB&M, and Mini-BESTest. Our results are in line withprevious findings of a high correla-tion among the BESTest and BBS(r�.87) and the Mini-BESTest(r�.96) in patients with PD.34,39

Although all of the clinical balancemeasures observed in this study canbe used to assess balance impair-ments in patients with subacutestroke, this study demonstrated thatthe BESTest may be more desirablethan other balance scales due to itslack of floor and ceiling effects. TheBESTest was shown to be suitable forpatients with subacute stroke acrossmany levels of functional ability, asevidenced by the distribution ofBESTest scores that covered thewhole range of functional ability. In

contrast, a significant floor effectwas found for the Mini-BESTest, par-ticularly in patients with low func-tional ability, and a significant flooreffect was present in all patientswith subacute stroke using theCB&M. The Mini-BESTest includeddynamic balance in standing andwalking, but most patients in the LFA

group had not yet regained their abil-ity to stand or walk independently;thus, they received scores of 0 onthese items of the Mini-BESTest,resulting in a floor effect (34.3%).However, no floor effect has beenreported when using the Mini-BESTest in community-dwelling peo-ple with chronic stroke, indicating

Figure 3.Receiver operating characteristic curve plots of the Balance Evaluation System Test(BESTest), Mini-Balance Evaluation System Test (Mini-BESTest), and Berg Balance Scale(BBS) for classifying patients with subacute stroke in the low and high functional abilitygroups. Arrow indicates the cutoff point for group classification.

Table 3.Cutoff Points With Associated Sensitivity, Specificity, Likelihood Ratios, and Area Under the Curve (AUC) of the ReceiverOperating Characteristic Curve Plot for the Balance Evaluation Systems Test (BESTest), Mini-Balance Evaluation Systems Test(Mini-BESTest), and Berg Balance Scale (BBS)a

Distinguishing HFA From LFA BESTest Mini-BESTest BBS

Cutoff point �49% �9 �19

AUC (95% CI) 0.88 (0.80, 0.95) 0.85 (0.77, 0.93) 0.85 (0.78, 0.93)

Sensitivity (95% CI) 71.4 (54.7, 83.7) 71.4 (54.7, 83.7) 85.7 (70.0, 94.1)

Specificity (95% CI) 91.4 (76.7, 97.7) 85.7 (70.0, 94.1) 68.6 (51.9, 81.5)

LR� (95% CI) 8.33 (2.77, 25.09) 5.00 (2.16, 11.56) 2.73 (1.64, 4.53)

LR� (95% CI) 0.31 (0.18, 0.53) 0.21 (0.09, 0.48) 0.33 (0.19, 0.57)

Accuracy 0.81 0.79 0.77

a HFA�high functional ability, LFA�low functional ability, 95% CI�95% confidence interval, LR��positive likelihood ratio, LR��negative likelihood ratio.

Validation of BESTest in Stroke

November 2014 Volume 94 Number 11 Physical Therapy f 9 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

that this assessment scale is moreappropriate for community-dwellingpeople poststroke.21 Similarly, theCB&M has been recommended forambulatory patients with stroke whohave moderate to mild poststrokedeficits,40 and our results supportedthat the assessment tasks in theCB&M were too difficult for thepatients with low functional ability.

In this study, the calculation of theceiling effect was performed 2 ways:by use of maximum score and by useof the score within top 10% of test(responsive ceiling effect). Ourresults are in agreement with Leddyet al,34 who found that the use of thescore within the top 10% would givemore information about the ability ofthe scale to measure balance prog-ress over time rather than using onlythe maximum score. Our resultsdemonstrated the responsive ceilingeffect of the BBS (45.7%) and PASS(60%) in the HFA group, suggestingthat the BBS and PASS may have lim-ited ability to detect balance prog-ress over time, which is very impor-tant for evaluating the effectivenessof balance rehabilitation in patientswith high functional ability.40 TheBBS and PASS had scores within thetop 10% of 51 to 56 and 34 to 36,respectively, and the scores mustchange at least 5.8 points41 on theBBS and by 2.2 points29 on the PASSto indicate a true change that is notdue to measurement error. There-fore, the BBS and PASS would not beproper to administer in patients withhigh functional stroke, as they haveinsufficient room to address true bal-ance progression.

The BESTest was developed to pro-vide information on which particularbalance systems were underlying bal-ance impairments. Results from ourstudy confirmed that balance deficitsin patients with subacute stroke cov-ered impairments of all postural con-trol subsystems represented in theBESTest (Tab. 2), but the amount of

subsystem deficit varied dependingon the level of motor impairment.For example, patients with LFA haddeficits in all postural control sys-tems (mean score of the BESTest wasless than 50%), except the stabilitylimit and verticality subsystems. Incontrast, patients in the HFA groupshowed more deficits in stability ingait. Therefore, the use of theBESTest in patients with subacutestroke can guide clinicians to tailortheir intervention to impairments ofspecific postural control subsystems.Although we did not study theresponsiveness of the BESTest, thedifferences in the item scoresbetween the LFA and HFA groupsindicated the potential of theBESTest to detect change in perfor-mance. All items in the BESTest, withthe exception of base of support andsitting verticality (left and right),showed high median scores and highpercentage of maximum scores inthe HFA group compared with theLFA group. Our participants receiveda daily rehabilitation program thatincluded interventions to maintainjoint range of motion and preventdeformity. Therefore, deformity orpain in the feet was rare in our par-ticipants, resulting in an initially highscore on the base of support item.Inaccurate representation of vertical-ity is often found in patients withstroke who have neglect or pushersyndrome.42,43 However, none ofour participants had been diagnosedwith either neglect or pusher syn-drome, and they all had perfectscores on the verticality items.

In this study, we also questionedwhether the scores from the BESTest,Mini-BESTest, and BBS could classifypatients with a high versus low levelof functional ability as measuredwith the FM-motor. This study dem-onstrated that the scores from all3 balance scales could be used toclassify patients with stroke into 2functional ability groups. However,the slightly larger LR� values sug-

gest the BESTest may be preferableto the BBS and Mini-BESTest for func-tional classification. Although thetime for administering the BESTest(30–35 minutes) is twice as long asfor the Mini-BESTest or the BBS(10–15 minutes), the informationgathered with the BESTest is moreuseful compared with the BBS andMini-BESTest in customizing balanceintervention and more sensitive thanthe Mini-BESTest to detect changesin balance performance. However,we question whether some items inthe BESTest that had maximumscores in the first assessment mayhave a floor effect and could limititem responsiveness in patients withsubacute stroke. Therefore, furtheranalysis, such as Rasch analysis, maybe required on BESTest data frompatients with subacute stroke toeliminate the least sensitive andredundant items, which will shortenthe assessment time. Future studiesabout sensitivity to detect changeand minimal detectable changes ofthe BESTest also are essential forevaluating the effectiveness of treat-ment programs. Furthermore, ourstudy was carried out only inpatients with early stroke; thus, theresults cannot be generalized topatients with chronic stroke.

In conclusion, the BESTest is a reli-able, valid, and comprehensive mea-sure of balance that can be usedin people with subacute strokethroughout different levels of func-tional ability without floor and ceil-ing effects.

Ms Chinsongkram, Dr Viriyatharakij, DrSaengsirisuwan, Dr Horak, and Dr Boonsin-sukh provided concept/idea/researchdesign. Ms Chinsongkram, Dr Horak, and DrBoonsinsukh provided writing. Ms Chin-songkram, Ms Chaikeeree, and Dr Saeng-sirisuwan provided data collection. Ms Chin-songkram and Dr Boonsinsukh provideddata analysis. Dr Boonsinsukh provided proj-ect management and fund procurement.Ms Chaikeeree and Dr Boonsinsukh pro-vided facilities/equipment. Dr Viriyatharakij

Validation of BESTest in Stroke

10 f Physical Therapy Volume 94 Number 11 November 2014 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

and Dr Horak provided consultation (includ-ing review of manuscript before submission).

This study was supported by the ThailandResearch Fund, the Office of Higher Educa-tion Commission, and Srinakharinwirot Uni-versity (grant no. RSA5580002).

DOI: 10.2522/ptj.20130558

References1 Mancini M, Horak FB. The relevance of

clinical balance assessment tools to differ-entiate balance deficits. Eur J Phys Reha-bil Med. 2010;46:239–248.

2 deHaart M, Geurts AC, Huidekoper SC,et al. Recovery of standing balance in post-acute stroke patients: a rehabilitationcohort study. Arch Phys Med Rehabil.2004;85:886–895.

3 Mizrahi J, Solzi P, Ring H, Nisell R. Posturalstability in stroke patients: vectorialexpression of asymmetry, sway activityand relative sequence of reactive forces.Med Biol Eng Comput. 1989;27:181–190.

4 Tyson SF, Hanley M, Chillala J, et al. Bal-ance disability after stroke. Phys Ther.2006;86:30–38.

5 Bonan IV, Colle FM, Guichard JP, et al.Reliance on visual information afterstroke, part I: balance on dynamic postu-rography. Arch Phys Med Rehabil. 2004;85:268–273.

6 Di Fabio RP, Badke MB. Stance durationunder sensory conflict conditions inpatients with hemiplegia. Arch Phys MedRehabil. 1991;72:292–295.

7 Shumway-Cook A, Woollacott MH. MotorControl: Translating Research to Prac-tice. 4th ed. Philadelphia, PA: LippincottWilliams & Wilkins; 2012.

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

9 Garland SJ, Stevenson TJ, Ivanova T. Pos-tural responses to unilateral arm perturba-tion in young, elderly, and hemiplegic sub-jects. Arch Phys Med Rehabil. 1997;78:1072–1077.

10 Marigold DS, Eng JJ. Altered timing of pos-tural reflexes contributes to falling in per-sons with chronic stroke. Exp Brain Res.2006;171:459–468.

11 Brown LA, Sleik RJ, Winder TR. Atten-tional demands for static postural controlafter stroke. Arch Phys Med Rehabil.2002;83:1732–1735.

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

13 Blum L, Korner-Bitensky N. Usefulness ofthe Berg Balance Scale in stroke rehabili-tation: a systematic review. Phys Ther.2008;88:559–566.

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

15 Franchignoni F, Horak F, Godi M, et al.Using psychometric techniques toimprove the Balance Evaluation SystemsTest: the mini-BESTest. J Rehabil Med.2010;42:323–331.

16 King LA, Priest KC, Salarian A, et al. Com-paring the Mini-BESTest with the Berg Bal-ance Scale to evaluate balance disorders inParkinson’s disease. Parkinsons Dis.2012;2012:375419. 2011 Oct 4 [Epubahead of print]. doi: 10.1155/2012/375419.

17 Padgett PK, Jacobs JV, Kasser SL. Is theBESTest at its best? A suggested brief ver-sion based on interrater reliability, validity,internal consistency, and theoretical con-struct. Phys Ther. 2012;92:1197–1207.

18 Godi M, Franchignoni F, Caligari M, et al.Comparison of reliability, validity, andresponsiveness of the mini-BESTest andBerg Balance Scale in patients with bal-ance disorders. Phys Ther. 2013;93:158–167.

19 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. 2011 Nov 30 [Epub ahead ofprint]. doi: 10.1155/2012/237673.

20 Duncan RP, Leddy AL, Cavanaugh JT, et al.Comparative utility of the BESTest, Mini-BESTest, and Brief-BESTest for predictingfalls in individuals with Parkinson disease:a cohort study. Phys Ther. 2013;93:542–550.

21 Tsang CS, Liao LR, Chung RC, Pang MY.Psychometric properties of the Mini-Balance Evaluation Systems Test (Mini-BESTest) in community-dwelling individu-als with chronic stroke. Phys Ther. 2013;93:1102–1115.

22 Sullivan KJ. On “Modified constraint-induced therapy. . .” Phys Ther. 2007;87:1560.

23 Dick JP, Guiloff RJ, Stewart A, et al. Mini-Mental State Examination in neurologicalpatients. J Neurol Neurosurg Psychiatry.1984;47:496–499.

24 Tombaugh TN, McIntyre NJ. The Mini-Mental State Examination: a comprehen-sive review. J Am Geriatr Soc. 1992;40:922–935.

25 Shewan CM, Kertesz A. Reliability andvalidity characteristics of the WesternAphasia Battery (WAB). J Speech Hear Dis-ord. 1980;45:308–324.

26 Gladstone DJ, Danells CJ, Black SE. TheFugl-Meyer assessment of motor recoveryafter stroke: a critical review of its mea-surement properties. Neurorehabil Neu-ral Repair. 2002;16:232–240.

27 Duncan PW, Goldstein LB, Matchar D,et al. Measurement of motor recoveryafter stroke outcome assessment and sam-ple size requirements. Stroke. 1992;23:1084–1089.

28 Berg KO, Maki B, Williams JI, et al. Clinicaland laboratory measures of postural bal-ance in an elderly population. Arch PhysMed Rehabil. 1992;73:1073–1080.

29 Chien CW, Hu MH, Tang PF, et al. A com-parison of psychometric properties of thesmart balance master system and the pos-tural assessment scale for stroke in peoplewho have had mild stroke. Arch Phys MedRehabil. 2007;88:374–380.

30 Howe JA, Inness EL, Venturini A, et al. TheCommunity Balance and Mobility Scale: abalance measure for individuals with trau-matic brain injury. Clin Rehabil. 2006;20:885–895.

31 Jorstad EC, Hauer K, Becker C, et al. Mea-suring the psychological outcomes of fall-ing: a systematic review. J Am Geriatr Soc.2005;53:501–510.

32 de Oliveira CB, de Medeiros IR, Frota NA,et al. Balance control in hemipareticstroke patients: main tools for evaluation. JRehabil Res Dev. 2008;45:1215–1226.

33 Cohen J. Statistical Power Analysis forthe Behavioral Sciences. 2nd ed. Hillsdale,NJ: Lawrence Erlbaum Associates; 1988.

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

35 Portney LG, Watkins MP. Foundations ofClinical Research: Applications to Prac-tice. 3rd ed. Norwalk, CT: Appleton &Lange; 1993.

36 Andresen EM. Criteria for assessing thetools of disability outcomes research. ArchPhys Med Rehabil. 2000;81:S15–S20.

37 McHorney CA, Ware JE Jr, Lu JF, Sher-bourne CD. The MOS 36-item Short-FormHealth Survey (SF-36), III: tests of dataquality, scaling assumptions, and reliabil-ity across diverse patient groups. MedCare. 1994;32:40–66.

38 Akobeng AK. Understanding diagnosistests 2: likelihood ratios, pre- and post-testprobabilities and their use in clinical prac-tice. Acta Paediatr. 2007;96:487–491.

39 Leddy AL, Crowner BE, Earhart GM. Utilityof the Mini-BESTest, BESTest, and BESTestsections for balance assessments in indi-viduals with Parkinson disease. J NeurolPhys Ther. 2011;35:90–97.

40 Knorr S, Brouwer B, Garland SJ. Validity ofthe Community Balance and Mobility Scalein community-dwelling persons afterstroke. Arch Phys Med Rehabil. 2010;91:890–896.

41 Stevenson TJ. Detecting change inpatients with stroke using the Berg Bal-ance Scale. Aust J Physiother. 2001;47:29–38.

42 Ting DSJ, Pollock A, Dutton GN, et al.Visual neglect following stroke: currentconcepts and future focus. Surv Ophthal-mol. 2011;56:114–134.

43 Abe H, Kondo T, Oouchida Y, et al. Prev-alence and length of recovery of Pushersyndrome based on cerebral hemisphericlesion side in patients with acute stroke.Stroke. 2012;43:1654–1656.

Validation of BESTest in Stroke

November 2014 Volume 94 Number 11 Physical Therapy f 11 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

Eligible Participants

(n=113)

Excluded Participants

(n=43)

Did not require continuedrehabilitation (n=30)

Cognitive impairment or severemotor aphasia (n=5)

Severe internal carotid arteryocclusion (n=2)

Musculoskeletal problem (n=6)

Included Participants

(n=70)

eFigure.Flowchart diagram of number of eligible participants, included participants, andexcluded participants with reason for exclusion.

Validation of BESTest in Stroke

November 2014 (eFigure, Chinsongkram et al) Volume 94 Number 11 Physical Therapy f 1 by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from

doi: 10.2522/ptj.20130558Originally published online June 12, 2014

Published online June 12, 2014PHYS THER. Rumpa BoonsinsukhSaengsirisuwan, Nitaya Viriyatharakij, Fay B. Horak and Butsara Chinsongkram, Nithinun Chaikeeree, VitoonStrokeSystems Test (BESTest) in People With Subacute Reliability and Validity of the Balance Evaluation

Information Subscription http://ptjournal.apta.org/subscriptions/

Permissions and Reprints http://ptjournal.apta.org/site/misc/terms.xhtml

Information for Authors http://ptjournal.apta.org/site/misc/ifora.xhtml

by Kristi DeShazo on September 15, 2014http://ptjournal.apta.org/Downloaded from