pediatric arm function test

Cerebral Palsy

ORIGINAL RESEARCH ARTICLE

Pediatric Arm Function TestReliability and Validity for Assessing More-Affected ArmMotor Capacity in Children with Cerebral Palsy

ABSTRACT

Uswatte G, Taub E, Griffin A, Rowe J, Vogtle L, Barman J: Pediatric arm function

test: Reliability and validity for assessing more-affected arm motor capacity in

children with cerebral palsy. Am J Phys Med Rehabil 2012;91:1060Y1069.

Objective: Although there are several validated upper-extremity measures in

young children with cerebral palsy, none of these primarily assess the capacity to

carry out actions and tasks with the more-affected arm. To address this need, we

developed the Pediatric Arm Function Test (PAFT), which involves the behavioral

observation of how children use their more-affected arm during structured play in a

laboratory or clinic. This article evaluates the reliability and validity of the PAFT

Functional Ability scale.

Design: In study 1, a total of 20 children between 2 and 8 yrs old with a wide

range of upper-extremity hemiparesis caused by cerebral palsy completed the

PAFT on two occasions separated by 3 wks. In study 2, a total of 41 children

between 2 and 6 yrs old with similar characteristics completed the PAFT and

received a grade reflecting the severity of more-affected arm motor impairment.

Results: In study 1, the PAFT test-retest reliability correlation coefficient was

0.74. In study 2, convergent validity was supported by a strong inverse correlation

(r = j0.6, P G 0.001) between the PAFT scores and the grade of impairment.

Conclusions: The PAFT Functional Ability scale is a reliable and valid mea-

sure of more-affected arm motor capacity in children between 2 and 6 yrs old with

cerebral palsy. It can be used to measure upper-extremity neurorehabilitation

outcome.

Key Words: Upper Extremity, Function, Assessment, Cerebral Palsy

Authors:Gitendra Uswatte, PhDEdward Taub, PhDAngi Griffin, MA, OTLJan Rowe, DrOTLaura Vogtle, PhD, OTR/LJoydip Barman, MS

Affiliations:From the Departments ofPsychology (GU, ET, JB), PhysicalTherapy (GU), and OccupationalTherapy (JR, LV), University of Alabamaat Birmingham (UAB); and theDepartment of Physical Therapy andOccupational Therapy, Children’s ofAlabama, Birmingham (AG).

Correspondence:All correspondence and requests forreprints should be addressed to:Gitendra Uswatte, PhD, Department ofPsychology, University of Alabama atBirmingham, 1530 3rd Ave South,Room CH415, Birmingham, AL, 35294.

Disclosures:Financial disclosure statements havebeen obtained, and no conflicts ofinterest have been reported by theauthors or by any individuals in controlof the content of this article.Supported by Grant HD040692 fromthe National Center for MedicalRehabilitation Research of NationalInstitute of Child Health & HumanDevelopment and the Department ofPsychology at University of Alabamaat Birmingham.

0894-9115/12/9112-1060/0American Journal of PhysicalMedicine & RehabilitationCopyright * 2012 by LippincottWilliams & Wilkins

DOI: 10.1097/PHM.0b013e318269ec76

1060 Am. J. Phys. Med. Rehabil. & Vol. 91, No. 12, December 2012

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Approximately one third of children with hemi-paresis caused by cerebral palsy (CP) exhibit motordeficits in their more-affected arm.1,2 Althoughseveral measures of upper-extremity function inyoung children with CP are available,3 there are novalidated instruments that primarily assess thecapacity to complete actions and tasks with themore-affected arm. The Pediatric Arm FunctionTest (PAFT)4 was developed to evaluate this aspectof arm function in 2- to 12-yr-olds.

The International Classification of Function5

outlines three domains of function for assessment:anatomy and physiology (i.e., body function andstructure), execution of actions and tasks (i.e., ac-tivity), and fulfilling social roles (i.e., participation).Furthermore, within the domain of activity, theInternational Classification of Function stresses sep-arately evaluating capacity, what a person can dowhen tested in a standardized environment with-out assistance from others or devices, and perfor-mance, what a person actually does in his/hereveryday environment.

In adult neurorehabilitation, a strong empha-sis has been placed in the last decade on measuringlimitations in activity, as opposed to impairmentsin body structure and function (e.g., restricted rangeof motion or spasticity). This has been spurred byconsiderably more empirical support for therapeuticapproaches that emphasize training on tasks, suchas Constraint-Induced Movement therapy (CIMT),6,7

than for approaches that emphasize training move-ments.8 Parallel developments in pediatric neuro-rehabilitation9 have led to a similar emphasis onmeasuring limitations in activity, that is, deficitsin executing actions and tasks.

A recent review of measures of this domainin children with congenital hemiplegia identifiesthree reliable and valid tests of more-affected armmotor capacity.3 However, only a minority of itemson these tests assess limitations in capacity in theactivity domain: the Melbourne Assessment of Uni-lateral Upper Extremity Function,10 46%3; theQuality of Upper Extremity Skills Test,11 35%3; andthe Shriners Hospital Upper Extremity Evaluation(SHUEE) dynamic positional assessment compo-nent,12 0%. Most of the items assess impairmentin the body structure and the function domain.The Shriners Hospital Upper Extremity Evaluationspontaneous functional assessment component,12 theAssisting Hand Assessment,13 and the ABILHAND-Kids14 have evidence of reliability and validity and apreponderance of items that assess limitations in theactivity domain (981%).3 These tests, however, assess

motor performance, that is, what children actually doin daily life, rather than capacity. Other widely usedtests that have upper-extremity components measureperformance in the activity domain regardless ofwhich arm is used (e.g., the Pediatric Evaluation ofDisability Inventory functional skills scale)15 or assesshowmuch assistance is needed from others (e.g., thePediatric Evaluation of Disability Inventory caregiverassistance scale,15 the Functional IndependenceMeasure for Children [WeeFIM]16).

Thus, there are no validated instruments thatprimarily measure the capacity of young childrenwith CP to complete actions and tasks indepen-dently with the more-affected arm. Such an in-strument would permit researchers to test whethernew and accepted interventions for the upper ex-tremity in children with CP produce changes in thiscritical domain and to examine the contribution ofthe changes in this domain to changes in everydayperformance and participation in social roles. Clin-icians who care for children with CP would gain atool for monitoring progress on building capacityto complete actions and tasks with the more-affectedarm and for identifying strengths and weaknesses inthis domain for the purpose of treatment planning.

To address this need, this article presents twostudies examining the measurement characteristicsof the PAFT Functional Ability scale. The PAFT is abehavioral observation system set in a laboratoryor clinic that contains both unilateral and bilat-eral upper-extremity actions and tasks in the con-text of structured play. The Functional Ability scaleis used by masked raters, who score the PAFT from avideo, to quantify how effectively a child uses his/her more-affected arm to carry out each item onthe test. Study 1 evaluates the test-retest reliabilityand the stability of the PAFT Functional Ability scaleafter removing items deemed poor by an itemanalysis. Study 2 evaluates the convergent validity ofthe streamlined test against an index of the sever-ity of more-affected arm motor impairment in aseparate sample. Study 2 also determines the re-sponsiveness of the streamlined test to change inmore-affected arm motor capacity.

STUDY 1: METHODS

ParticipantsTwenty-one children between 2 and 8 yrs old

with a wide range of severity of upper-extremityhemiparesis caused by CP were enrolled. Table 1lists the eligibility criteria and the recruitment pro-cedures. Table 2 summarizes the participants’ char-acteristics. In both this study and study 2, informed

www.ajpmr.com Arm Function Test in Cerebral Palsy 1061


TABLE 1 Inclusion and exclusion criteria and recruitment procedures for studies 1 and 2

Study 1 Study 2

Inclusion criteria Between 2 and 8 yrs old, clinical diagnosisof CP, upper-extremity hemiparesis

Between 2 and 6 yrs old; stroke in theprenatal, perinatal, or very earlyantenatal period confirmed by magneticresonance imaging; upper-extremityhemiparesis; substantial nonuse of themore-affected upper extremity in dailylife (i.e., PMAL arm use score of e2.5)

Exclusion criteria Serious or recurring medical complicationsthat would interfere with studyparticipation (e.g., uncontrolled seizures)

Serious or recurring medicalcomplications that would interfere withstudy participation (e.g., uncontrolledseizures), spasticity medication withinthe last 3 mos, previous pediatric CIMT,fixed contractures in the more-affectedupper extremity (Ashworth score of 94)

Recruitment procedures The participants were recruited from apediatric outpatient physicalrehabilitation facility at an urban medicalcenter in the southeastern United States.They were paid $50 for theirparticipation. Consecutive children whomet the eligibility criteria were enrolled.

The families of the children eithercontacted our laboratory because oftheir interest in CIMT or were referredby local rehabilitation professionals.Consecutive children who met theeligibility criteria were enrolled.

CP, cerebral palsy; PMAL, Pediatric Motor Activity Log21; Ashworth, Modified Ashworth Scale35; CIMT, Constraint-InducedMovement therapy.

TABLE 2 Demographic and more-affected arm characteristics of participants

CharacteristicsStudy 1(n = 21)

Study 2(n = 41)

Age, yrsMean T SD 4.9 T 1.8 3.7 T 1.5Range 2Y8 2Y6

Female, n (%) 13 (62) 25 (61)Race, n (%)

European American 15 (71) 35 (85)African American 4 (19) 4 (10)Other 2 (10) 2 (5)

Paresis of right side, n (%) 16 (76) 27 (66)Severity of more-affected arm motor impairment,a n (%)

Grade 2 NAb 13 (32)Grade 3 NA 13 (32)Grade 4 NA 11 (27)Grade 5 NA 4 (9)

Everyday ability to handle objects regardless of arm used, MACS level, n (%)Level I (with ease and success) 5 (24) 7 (17)Level II (with reduced quality and/or speed but mostly successful) 10 (48) 16 (39)Level III (with difficulty and limited success, needs help with setup) 6 (28) 18 (44)

More-affected arm capacity for completing tasks, PAFT Functional Ability scale, mean T SD points (95% CI)Total score 2.7 T 0.7 (2.4Y3.0) 2.2 T 0.7 (2.0Y2.4)Unilateral subtotal score 2.7 T 0.7 (2.4Y3.0) 2.3 T 0.8 (2.0Y2.6)Bilateral subtotal score 2.7 T 0.8 (2.3Y3.1) 2.2 T 0.7 (2.0Y2.4)aThe testers assigned the participants a Bgrade[ based on active range of motion available at the upper-extremity joints. Table 5

describes the scheme used.bThese data were not collected for study 1.MACS, Manual Ability Classification System; PAFT, Pediatric Arm Function Test; CI, confidence interval; NA, not applicable.

1062 Uswatte et al. Am. J. Phys. Med. Rehabil. & Vol. 91, No. 12, December 2012


consent was received from the legal guardians of allparticipants; the study procedures were approved byour institutional review board.

MeasuresThe PAFT is a behavioral observation system

that quantifies more-affected arm function of chil-dren with upper-extremity hemiparesis from a videoof structured play in the treatment setting. It is asubstantially modified homologue of the ToddlerArm Use Test.17 The Toddler Arm Use Test itemswere generated by identifying unimanual activi-ties commonly undertaken by 2- to 8-yr-olds andselecting 22 activities so that distal and proximalarm movement and different types of grasp weredemanded.17 Based on experience using the ToddlerArm Use Test to evaluate pediatric CIMT,17 biman-ual activities were added when developing thePAFT. In addition, the variety of unilateral activi-ties was increased to demand arm movement indifferent planes and during different postures, forexample, sitting vs. standing.4 Because the item-rating scheme for the Toddler Arm Use Test provedcumbersome, the Willingness and Amount of Par-ticipation scales were dropped, and the How Wellscale was replaced with the Functional Ability scalefrom the Wolf Motor Function Test.18,19 This scale,however, was modified so that appropriate use ofboth arms during bimanual activities, which areabsent from the Wolf Motor Function Test, wouldnot be penalized.

Table 3 lists the 17 unilateral and 9 bilateralupper-extremity items embedded within a PAFTplay session. All 26 items are presented first in afixed order, with no cues given about which arm touse. On item 1, for example, a tester positions a ball15.2 cm directly above children’s heads and saysBTouch the ball.[ The tester then revisits any itemsthat children do not attempt with the more-affectedarm. The tester now explicitly asks children to usethe more-affected arm for the unilateral items andboth arms for the bilateral items. If children still donot attempt to use the more-affected arm, the testerasks children to carry out the item again with thesame directive to use the more-affected arm. Thetester is now allowed to physically restrain the less-affected arm.

After the testing session, trained, masked ratersevaluate more-affected arm motor capacity item byitem from a video using the Functional Abilityscale (Table 4). The Functional Ability subtotalscore for the unimanual items is the sum of thefirst-administration scores from items 1Y17 that are

only administered once plus the sum of the final-administration scores from items 1Y17 that areadministered more than once, all divided by thenumber actually tested for items 1Y17. The subto-tal for the bilateral items is calculated in the sameway for items 18Y26. The Functional Ability totalscore is the average of the unilateral and bilateralsubtotal scores. This quantity was chosen, ratherthan the average of all items, to give the scores fromthe unilateral and bilateral sections equal weight.

An index of more-affected arm motor perfor-mance is also derived from the raters’ scores. TheExtremity Preference total score is the percentageof the eight unilateral items (i.e., items 1Y4, 7, 12,15, and 16) that are attempted with the more-af-fected arm on the first administration (i.e., Func-tional Ability score 90). The other unilateral itemsare excluded because they have poor item-totalcorrelation or test-retest reliability coefficients (i.e.,values G0.3) with respect to this index.20 Only tasksfrom the unilateral portion of the test are countedbecause the bilateral activities implicitly constrainchildren to use the more-affected arm. Only thefirst-administration scores are counted becausechildren do not receive prompts about which arm touse. As such, the children’s spontaneous choicesabout whether to use their more-affected arm arethought to reflect how much they actually use thatarm in daily life. The reliability and validity of theExtremity Preference score are reported elsewhere.21

Instructions for administering, videotaping, and scor-ing the PAFT are available online.20

ProcedureThe PAFT was administered on two occasions

(test 1, test 2) separated by approximately 3 wks.Although no treatment was provided by the study,children were permitted to receive customary care,which ranged from no physical rehabilitation to2 hrs/wk of outpatient physical rehabilitation. Themaximum amount of upper-extremity rehabilita-tion was 1 hr/wk. Such treatment has been previ-ously shown not to affect more-affected arm motorcapacity in children with CP.4,17 The PAFT, whichtypically took 30 mins, was carried out by pairs ofpediatric occupational therapists. One administeredthe tasks while the other controlled the cameraand recorded the data. The testers had approxi-mately 7 hrs of training; they read the manual,shadowed an experienced tester, and conducted twoto three supervised tests. The testing was done ina private treatment room in an outpatient rehabili-tation clinic.



TABLE

3CharacteristicsofthePAFTitems(n

=20)

Dim

ension

sClassified

inCon

tent

Analysis

Item

Item

-Total

Correlation

bICFDom

ain(Code)

Fing

erMovem

ent

Required

Plane

ofArm

Movem

ent

Posture

inWhich

ActivityisDon

e

Unilateralitem

s1

Reach

abovethehead

0.42

Bodyfunction

(7101)

No

Overhead

Sitting

2Reach

atwaist

level

0.70

Bodyfunction

(7101)

No

Straight

ahead

Sitting

3Reach

across

midlin

e0.67

Bodyfunction

(7101)

No

Across

midlin

eSitting

4Grasp

ball

0.53

Activity

(4452,

4450)

Yes

Straight

ahead

Sitting

5Carry

ball

0.84

Activity

(4401,

4301)

No

Straight

ahead

Standing

6Release

ballinto

cup

0.82

Activity

(4403)

Yes

Straight

ahead

Standing

7Po

urballou

tof

cup

0.80

Activity

(4452,

4453)

No

Totheside

Standing

8Th

rowballon

totarget

0.74

Activity

(4454)

Yes

Overhead

Standing

9Wavebye-byea

0.24

Activity

(3350)

Noc

Totheside

oroverhead

Standing

10Protective

extensionin

sittinga

0.51

Bodyfunction

(755)

No

Totheflo

orSitting

11Isolated

fingeruse

0.65

Activity

(4452,

4402)

Yes

Straight

ahead

Sitting

12Rem

ovebig-kn

obpu

zzle

piece

0.83

Activity

(4452,

4400)

Yes

Straight

ahead

Sitting

13Crayongrasp

0.76

Activity

(4452,

4400)

Yes

Straight

ahead

Sitting

14Crayonuse

0.82

Activity

(4401,

4402)

No

Across

midlin

eSitting

15Grasp

cracker-sizedfood

(e.g.animalor

graham

cracker,saltine)

0.89

Activity

(4452,

4400)

Yes

Straight

ahead

Sitting

16Grasp

smallfood

item

(e.g.cheerio

orraisin)

0.83

Activity

(4452,

4453,4

450)

Yes

Straight

ahead

Sitting

17Eat

withaspoon

0.83

Activity

(4452,

4453,4

450)

Yes

Straight

ahead

Sitting

Bilateralitem

s18

Separate

pull-aparttoy

0.84

Activity

(4452,

4450)

Yes

Totheside

Sitting

19Carry

largeball(e.g.b

asketball)

0.69

Activity(410,4452,4300,4301)

Noc

Totheflo

orStanding

20Th

rowballinto

hoop

0.75

Activity

(415,4

454)

Yes

Straight

ahead

Standing

21Placehaton

thehead

0.72

Activity

(4452,

4300,4

450)

Yes

Overhead

Sitting

22Pu

ton

boots(using

hand

s)0.39

Activity

(4452,

4300,4

450)

Yes

Straight

aheador

totheflo

orc

Sitting

23Com

eto

sita

0.10

Activity

(410)

No

Totheflo

orNA

24Quadrup

edweigh

t-bearing

0.47

Activity

(415)

Noc

Totheflo

orQuadrup

ed25

Crawlin

g0.64

Activity

(455)

Noc

Totheflo

orQuadrup

ed26

Com

eto

standusingbencha

0.30

Activity

(410)

No

Straight

ahead

NA

Forthecontentanalysis,twopairsof

pediatricOTs

independ

entlyclassifiedeach

item

alon

gfivedimension

s.Fo

llowingthemetho

ddescribedby

Cieza

etal.,3

6thepairsbrokeup

theitem

into

thecomponent

action

sbeingtested

andassign

edan

ICFcode

toeach

component.T

heICFdomaincolumnreportswhether

theitem

was

placed

inthebody

function

sandstructure(bodyfunction

)or

intheactivities

andparticipation(activity)

domainandreportsthecodesassign

edto

each

component

action

.The

key3

7to

thesecodesisgivenin

thelistof

abbreviation

sforthistable.

Thepairsjudged

whether

theitem

requ

ired

theuseof

thefin

gers

tobe

done

successfully.38

Thepairscodedtheplaneof

movem

entin

which

reaching

was

requ

ired.

Thepairscodedtheposturein

which

theitem

was

done.

Thepairsjudged

whether

theitem

was

ageappropriate.Because

allitem

sweredeem

edso,these

rating

sareno

treported

inthistable.

Excluding

item

s9,10,23,and26,w

hich

weredroppedon

thebasisoftheitem

analysis,agreementb

etweenthepairsofratersforclassificationofitem

salon

gdimension

1,thatis.,body

function

vs.activity,was

100%

.For

dimension

s2,3,4,

and5,

agreem

entwas

86%,9

5%,1

00%,and

100%

,respectively.

a The

aprioricriterionfordropping

anitem

was

anitem

-totalcorrelation2

5ofG0.3.Item

10,w

hich

testswhether

child

renusetheirmore-affected

arm

Breflexively[

tohelp

maintaintheirbalancewhenitisperturbed,was

droppedon

apost

hocbasisbecausearm

movem

entwas

elicited

rather

than

voluntary,as

onallthe

othertasks.Item

10isalso

moredifficultto

administerthan

aretheothers.

b For

theitem

analysis,the

Pearsoncorrelationbetweeneach

item

andsum

oftheotheritem

sisreported.

c Classification

ofitem

sforwhich

therewas

disagreementbetweenthepairsof

raters

was

adjudicatedby

authorsG.U

swatte,E

.Taub,

andA.

Griffin.

PAFT

,PediatricArm

Function

Test;ICF,

InternationalC

lassification

ofFu

nction

;OTs,occup

ationaltherapists;NA,

notapplicable;3350,producingbody

lang

uage;410,chang

ingbasicbody

position

;415,m

aintaining

abody

position

;4300,

lifting

;4301,

carrying

inthehand

s;4400,p

icking

up;4

401,

grasping

;4402,

manipulating;

4403,releasing

;4450,

pulling

;4452,

reaching

;4453,

turningor

twisting

thehand

sor

arms;4454,throw

ing;

455,

movingarou

nd;7

101,

mobility

ofa

sing

lejoint;755,

involuntarymovem

entreaction

function

s.



After the testing, the children’s more-affectedarm Functional Ability was scored from the video byone of three pediatric occupational therapists, all ofwhom were masked to the testing occasion. Eachrater had approximately 25 hrs of training; theyread the test manual, scored three PAFT sessionsjointly under the supervision of an experiencedtester, and scored eight sessions independently,followed by a joint review of each session under thesupervision of the experienced tester. Before scor-ing the study data, the raters demonstrated highinterrater reliability on an additional set of ten tests(intraclass correlation22 = 0.96).

To describe everyday manual ability, an occu-pational therapist assigned each child a ManualAbility Classification System (MACS)23 level basedon a retrospective review of the children’s records.Although the MACS is done ideally by professionaland family caregivers together, adequate agreementbetween professional and family MACS ratings(intraclass correlation, range, 0.73Y0.85) has beenreported even when up to 60% of the professionals’ratings were based solely on chart reviews.24

Data AnalysisItem and content analyses of the PAFT were

conducted. Subsequently, several statistics werecalculated to evaluate the reliability and validity ofthe Functional Ability test score. To index internalconsistency, Cronbach alpha was calculated usingthe test 1 data. Test-retest reliability was evaluatedby the intraclass correlation type22 2,1 between thetest 1 and test 2 scores. Power to detect adequatelevels of internal consistency and test-retest reli-ability, that is, coefficients25 greater than or equal to

0.7, was very good (90.93).26,27 Stability was eval-uated by paired t-tests of the test 1 and test 2 scores.Data from one participant who did not return forfollow-up testing without giving a reason was ex-cluded from all analyses.

STUDY 1: RESULTS

Item and Content AnalysisThe item analysis (Table 3) suggested that

items 9 (wave bye-bye), 23 (come to sit), and 26(come to stand using bench) be dropped. TheFunctional Ability scores for these items correlatedweakly with the summary score for the rest of thetest, indicating that these items do not assess thesame parameter as do the others. In addition, item10, which tests whether children use their more-affected arm Breflexively[ to help maintain theirbalance when it is perturbed, was dropped becausearm movement was elicited rather than voluntary,as on all the other tasks. Item 10 was also moredifficult to administer than were the others.

For the remaining items, the content analysis(Table 3) indicated that all items were develop-mentally appropriate and that all, except for items1Y3 (reach above the head, at waist level, and acrossmidline) were in the International Classification ofFunction activity domain. Finger movement wasnecessary for 59% of the items. The breakdown forplane of arm movement, in which reaching tookplace during each item, was straight ahead (59%),to the floor (18%), overhead (14%), midline (9%)and side (9%). The percentages add up to greaterthan 100 because item 22, putting on boots, can bedone by reaching either straight ahead or to the

TABLE 4 PAFT Functional Ability scale

Rating Anchor

0 Not used: does not attempt with the upper extremity being tested1 Very poor: affected upper extremity does not participate functionally; however, an attempt is made or the

less-affected upper extremity is used to move the upper extremity being tested. In bilateral tasks, themore-affected upper extremity serves as a helper but only through part of the task.

2 Poor: requires assistance of the less-affected upper extremity, another body part, or the therapist for minorreadjustments or change of position or requires more than two attempts to complete or accomplishesvery slowly. Movement is governed by synergy. In bilateral tasks, the more-affected upper extremityserves only as a helper.

3 Fair: a moderate amount of synergy is seen (i.e., synergistic pattern observed with some involuntary postureor movement and/or lack of control of movement, compensatory strategies with trunk/shoulder/elbowsobserved), or the task is performed somewhat slowly or with effort. In unilateral tasks, themore-affected upper extremity does not require assistance from the other upper extremity.

4 Good: movement is slightly slower; may lack precision, fine coordination, or fluidity; some synergy maybe present, but isolation of movement is predominant

5 Normal: movement seems to be normal

PAFT, Pediatric Arm Function Test.



floor. Sixty-four percent of the items were done insitting; 27%, in standing; and 9%, in quadrupedalposture. Only item 19 (carry large ball) required atransition from one posture to another, that is,squat to stand.

Reliability, Stability, and PrecisionAfter dropping items 9, 10, 23, and 26, the in-

ternal consistency of the PAFT Functional Abilitytotal score was high (Cronbach > = 0.96). Test-retest reliability was adequate (intraclass correla-tion = 0.74). Furthermore, the total scores did notchange from one administration to another (mean[SD] change, 0.1 [0.5]; confidence interval, j0.2 to0.3; P = 0.51).

For the unilateral and bilateral sections of thePAFT, the Cronbach > values were 0.96 and 0.87,respectively, whereas the test-retest reliabilitycoefficients were 0.7 and 0.68, respectively. As forthe total score, the subtotal scores for these sec-tions did not change from one administration tothe other (unilateral subtotal, mean [SD] change,0.0 [0.5]; confidence interval, j0.2 to 0.3; P = 0.76;

bilateral subtotal, mean [SD] change, 0.2 [0.6];confidence interval, j0.1 to 0.5; P = 0.29).

STUDY 2: METHODSParticipants

The participants were 41 children between 2and 6 yrs old with a wide range of severity of up-per Extremity hemiparesis caused by CP. Their datawere assembled from two clinical trials: (1) a ran-domized controlled trial testing the efficacy of pe-diatric CIMT and (2) a dose-response trial testingthe effect of reducing the duration of pediatric CIMTon the magnitude of treatment gains. The eligibil-ity criteria and the recruitment procedures forthese two trials were the same; they are describedin Table 1. Table 2 summarizes the participants’characteristics.

Measures, Procedure, and Data AnalysisIn the randomized controlled trial,4 ten chil-

dren received 97.5 hrs of CIMT for 3 wks: 90 hrswere for more-affected arm training and 7.5 hrs werefor a package of behavioral techniques4,7 designed

TABLE 5 Grading system for severity of more-affected arm motor impairment in children withcerebral palsy

Grade ofDeficit

Joint

Shoulder Elbow Wrist Fingers Thumb

2 (mild/moderate)

WNL/mildlimitationa

in flexion orabduction

WNL/mildlimitationin extension



WNL/mild limitationin lateral abduction

3 (moderate) Moderatelimitationb

in flexion orabduction

Moderatelimitationin extension



Moderate limitationin lateral abduction

4 (moderatelysevere)

Severelimitationc

in flexion orabduction, but930 degrees

Severe limitationin extension



Severe limitationin lateral abduction

5A (severe) e30 degreesflexion orabduction

Initiation offlexion orextension

Initiation of wrist, fingers or thumb movement

5B (verysevere)

e30 degreesflexion orabduction

Initiation offlexion orextension

No initiation of wrist, fingers and thumb movement

The movements described are the minimum motor criteria, that is, if a child does not meet the AROM criteria listed for thegrade at even one joint, he/she would be placed in the grade corresponding to the movement present at the worst joint.

aAROM is greater than two thirds to just below the reference range.bAROM is between one half and two thirds of the reference range.cAROM is less than one half of the reference range but can initiate movement.WNL, within normal limits; AROM, active range of motion.



to transfer training gains to daily life. The childrenalso wore a cast on their less-affected arm to limitits use. Twelve children received customary care.In the dose-response trial, nine children received52.5 hrs of CIMT for 3 wks, whereas ten childrenreceived 35 hrs of CIMT for 2 wks. Two reviewsconclude that the efficacy of pediatric CIMT issupported.28,29

The participants completed the PAFT beforeand after CIMT or customary care. The testersgave the participants a BGrade[ before treatmentthat reflected the severity of more-affected armmotor impairment.4 Table 5 describes this gradingscheme, which is based on active range of motionpresent at the upper-extremity joints. In addition,an occupational therapist assigned each child aMACS level, as in study 1. The PAFT administrationand scoring were conducted in the same manner asthat in study 1. The raters, however, were maskedto the group assignment and grade and the MACSlevel in addition to the testing occasion.

Convergent validity was evaluated by the Spear-man correlation between the PAFT Functional Abil-ity scores and the more-affected arm Grade beforetreatment. A strong inverse correlation (r e j0.5)27

was predicted on the basis of previous studies show-ing that the severity of upper-extremity motorimpairment is inversely related to the capacity tocarry out tasks.30Y32 Power to test this predictionwas very good (90.93).27 Responsiveness to changewas indexed by the standardized response mean27,33:the mean change in the Functional Ability scoresfrom pretreatment to posttreatment for childrenwho received CIMT in the randomized controlledtrial divided by the standard deviation of the cor-responding change scores in children in the controlgroup. For d ¶, which is a statistic similar to thestandardized response mean, values greater than orequal to 0.57 are considered large.27 Two controlswho withdrew because of medical complicationsafter pretreatment testing were excluded from thisanalysis.

STUDY 2: RESULTSThe correlation between the PAFT Functional

Ability total scores and grade was j0.6 (P G 0.001).The standardized response mean for the FunctionalAbility total score was 0.73. For both the FunctionalAbility unilateral and bilateral subtotal scores, thecorrelation with grade was the same as for the to-tal score. The standardized response mean valuesfor the unilateral and bilateral subtotal scores were0.55 and 0.53, respectively.

OVERALL DISCUSSIONAs noted in the BIntroduction,[ none of the

existing measures with evidence of validity for as-sessing more-affected arm function in young chil-dren with CP primarily assess the capacity toexecute actions and tasks. This International Clas-sification of Function domain, activity, is importantbecause of the considerable empirical support fortherapeutic approaches that emphasize training ontasks8,9 and the limits that the deficits in capacityto execute actions and tasks place on participation insocial roles.34 The content analysis in study 1 indi-cates that the PAFT Functional Ability scale, excludingthe four items that the item analysis deems poor, hasa preponderance of items that address this domain.

In addition, study 1 indicates that the PAFTFunctional Ability total score is a reliable and stablemeasure of more-affected arm capacity. The partici-pants, who received little or no upper-extremity re-habilitation between the testing occasions, had totalscores at the second occasion that were similar inrank and absolute value to those at the first occasion.

Study 2 indicates that the Functional Abilitytotal score validly measures more-affected arm ca-pacity. The total score and the severity of upper-extremity motor impairment were strongly in-versely correlated. Study 2 also shows that the totalscore is responsive to change. The mean gain inchildren who received 35 hrs or more of CIMT waslarge relative to the variability in total scores inchildren who received usual care, that is, little or noupper-extremity rehabilitation.

In studies 1 and 2, the reliability and validity ofthe Functional Ability subtotal scores for the uni-lateral and bilateral sections of the test are largelysimilar to those for the total score. The highCronbach alpha value for the Functional Abilityscores from the entire test suggests that the items,regardless of whether they are from the unilateralor the bilateral section, all assess a single underly-ing construct, which supports reporting the totalscore. Reporting only the total score is also sup-ported by the high correlations between the base-line unilateral and bilateral subtotal scores in bothstudies 1 and 2 (r’s Q 0.8, P’s G 0.001). The unilateraland bilateral section subtotal scores might be ofinterest in special cases, such as evaluating theeffects on bilateral function of an intervention thatonly trains more-affected arm function.

Study LimitationsLimitations are study samples too small to permit

subanalyses by age or sex, exclusion of participants



older than 6 yrs in study 2, and demonstration ofconvergent validity with the classification sys-tem used rather than an established measure. Withregard to age, the PAFT is thought to be appropriateup to 12 yrs. For older children, whose activities arenot that different from adults, the Wolf MotorFunction Test18,19 would be appropriate. With re-gard to the convergent measure, the MACS was notselected because it assesses how well a child handlesobjects regardless of which arm is used. Thus, theMACS ratings can reflect how well the child uses themore-affected arm, less-affected arm, or both incoordination.23 Given this indirect relationshipwith more-affected arm function, it was not sur-prising that the association of the PAFT with theMACS was significant (r =j0.44, P = 0.001) but lessstrong than with the classification system used asthe convergent measure, that is, the grade of more-affected arm impairment. Future studies mightcorrelate the PAFT with an established measure suchas the Melbourne Assessment of Unilateral UpperExtremity Function.10

OVERALL CONCLUSIONSStudies 1 and 2 support the reliability and va-

lidity of the PAFT Functional Ability scale for mea-suring the capacity of children between 2 and 6 yrsold with upper-extremity hemiparesis caused by CPto complete actions and tasks with their more-af-fected arm. Researchers might use the FunctionalAbility scale to quantify more-affected arm motorstatus in children with CP or other conditions withsimilar upper-extremity sequelae in the above-men-tioned age range. Collection of such data wouldpermit examination of the relationship between thecapacity to complete actions and tasks with themore-affected arm and the actual use of that capacity ineveryday life, for example. Researchersmight also usethe PAFT to document changes in motor status afterupper-extremity pediatric neurorehabilitation. Col-lection of this type of data would permit evaluationof the effect of new and accepted interventions forchildren with asymmetric upper-extremity motorimpairment on capacity to perform actions and taskswith the more-affected arm. In addition to doc-umenting changes after treatment, clinicians mightuse the PAFT to assist with treatment planning. Ac-tivities on which children show limited capacity be-fore treatment might be targeted for training.

ACKNOWLEDGMENTS

We thank Anna Ballenger, OTR/L; MargieanBurks, MS, OTR/L; Jane Colburn, MA, OTL;Katherine Goldman, MS, OTR/L; Elise Hollinhead,

MS, OTR/L; and Bryony Lane, MS, OTL, for carryingout the content analysis of the Pediatric ArmFunction Test.

REFERENCES

1. Green LB, Hurvitz EA: Cerebral palsy. Phys MedRehabil Clin N Am 2007;18:859Y82

2. Hagberg B, Hagberg G, Beckung E, et al: Changingpanorama of cerebral palsy in Sweden. VII. Pre-valance and origin in the birth year period 1991Y94.Acta Paediatr 2001;90:271Y7

3. Gilmore G, Sakzewski L, Boyd R: Upper-extremityactivity measures for 5- to 16-year old children withcongenital hemiplegia: A systematic review. Dev MedChild Neurol 2010;52:14Y21

4. Taub E, Griffin A, Uswatte G, et al: Treatment ofcongenital hemiparesis with pediatric Constraint-Induced Movement therapy. J Child Neurol 2011;26:1163Y73

5. World Health Organization: International Classifi-cation of Functioning, Disability, and Health. Ge-neva, World Health Organization, 2001

6. Taub E, Miller NE, Novack TA, et al: Technique toimprove chronic motor deficit after stroke. Arch PhysMed Rehabil 1993;74:347Y54

7. Taub E, Uswatte G, King DK, et al: A placebo-controlled trial of Constraint-Induced Movementtherapy for upper extremity after stroke. Stroke2006;37:1045Y9

8. Langhorne L, Coupar F, Pollack A: Motor recoveryafter stroke: A systematic review. Lancet Neurol2009;8:741Y54

9. Damiano D: Rehabilitative therapies in cerebral palsy:The good, the not as good, and the possible. J ChildNeurol 2009;24:1200Y4

10. Johnson LM, Randall MJ, Reddidough DS, et al: De-velopment of a clinical assessment of quality ofmovement for unilateral upper extremity function.Dev Med Child Neurol 1994;36:965Y73

11. DeMatteo C, Law M, Russell D, et al: Quality of UpperExtremity Skills Test. Hamilton, Canada, McMasterUniversity, Clinical Research Unit, 1992

12. Davids JR, Peace LC, Wagner LV, et al: Validation ofthe Shriners Hospital for Children Upper ExtremityEvaluation (SHUEE) for children with hemiplegiccerebral palsy. J Bone Joint Surg Am 2006;88:326Y33

13. Krumlinde-Sundholm L, Eliasson AC: Developmentof the Assisting Hand Assessment: A rasch-builtmeasure intended for children with unilateral upperextremity impairments. Scand J Occup Ther 2003;10:16Y26

14. Arnould C, Penta M, Renders A, et al: ABILHAND-Kids: A measure of manual ability in children withcerebral palsy. Neurology 2004;63:1045Y52

15. Haley SM, Coster WJ, Ludlow LH, et al: PediatricEvaluation of Disability Inventory (PEDI). Boston,MA, New England Medical Center Hospitals, 1992.



16. McCabe M, Granger C: Content validity of a pediat-ric functional independence measure. Appl Nurs Res1990;3:120Y2

17. Taub E, Ramey SL, DeLuca SC, et al: Efficacy ofConstraint-Induced Movement therapy for childrenwith cerebral palsy with asymmetric motor impair-ment. Pediatrics 2004;113:305Y12

18. Wolf SL, Lecraw DE, Barton LA, et al: Forced use ofhemiplegic upper extremities to reverse the effect oflearned nonuse among chronic stroke and head-injured patients. Exp Neurol 1989;104:125Y32

19. Morris DM, Uswatte G, Crago J, et al: The reliability ofthe Wolf Motor Function Test for assessing upperextremity motor function following stroke. Arch PhysMed Rehabil 2001;82:750Y5

20. Taub E, Griffin A, Uswatte G: Pediatric Arm Func-tion Test (PAFT) Manual 2012. Available at: http://www.uab.edu/citherapy/training-manuals-a-publica-tions. Accessed February 20, 2012

21. Uswatte G, Taub E, Griffin A, et al: The PediatricMotor Activity Log-Revised: Assessing real-world armuse in children with cerebral palsy. Rehabil Psychol2012;57:149Y58.

22. Shrout PE, Fleiss JL: Intraclass correlations: Usesin assessing rater reliability. Psychol Bull 1979;86:420Y8

23. Eliasson AC, Krumlinde-Sundholm L, Rosblad B,et al: The Manual Ability Classification System (MACS)for children with cerebral palsy: Scale developmentand evidence of reliability and validity. Dev Med ChildNeurol 2006;48:549Y54

24. Morris C, Kurinczuk JJ, Fitzpatrick R, et al: Reli-ability of the Manual Ability Classification System forchildren with cerebral palsy. Dev Med Child Neurol2006;48:950Y3

25. Kline P: Handbook of Psychological Testing, 2nd ed.New York, NY, Routledge, 2000

26. Bonnett DG: Sample size requirements for testingand estimating coefficient alpha. J Educ Behav Stat2002;27:335Y40

27. Cohen J: Statistical Power Analysis for the Behav-

ioral Sciences, 2nd ed. Hillsdale, NJ, LawrenceErlbaum Associates, 1988

28. Huang HH, Fetters L, Hale J, et al: Bound for success:A systematic review of Constraint-Induced Move-ment therapy in children with cerebral palsy supportsimproved arm and hand use. Phys Ther 2009;89:1126Y41

29. Brady K, Garcia T: Constraint-Induced Movementtherapy (CIMT): Pediatric applications. Dev DisabilRes Rev 2009;15:101Y11

30. Ostensjo S, Brogren Carlsber E, Vollestad NK: Motorimpairments in young children with cerebral palsy:Relationship to gross motor function and everydayactivities. Dev Med Child Neurol 2004;46:580Y9

31. Burridge J, Turk R, Notley SV, et al: The relationshipbetween upper extremity activity and impairmentin post-stroke hemiplegia. Disabil Rehabil 2009;31:109Y17

32. Wichers M, Hilberink S, Roebrueck ME, et al: Motorimpairments and activity limitations in children withspastic cerebral palsy: A Dutch population-basedstudy. J Rehabil Med 2009;41:367Y74

33. Guyatt GH, Walter S, Norman G: Measuring changeover time: Assessing the usefulness of evaluative in-struments. J Chronic Dis 1987;40:171Y8

34. van Meeteren JV, Roebrueck ME, Celen E, et al:Functional activities of the upper extremity of youngadults with cerebral palsy: A limiting factor for par-ticipation? Disabil Rehabil 2008;30:387Y95

35. Bohannon RW, Smith MB: Interrater reliability of aModified Ashworth Scale of muscle spasticity. PhysTher 1987;67:206Y7

36. Cieza A, Geyh S, Chatterji S, et al: ICF linking rules:An update based on lessons learned. J Rehabil Med2005;37:212Y8

37. WHO: ICF Browser 2012. Available at: http://apps.who.int/classifications/icfbrowser/. Accessed January15, 2012

38. Uswatte G, Taub E, Morris D, et al: The Motor ActivityLog-28: Assessing daily use of the hemiparetic armafter stroke. Neurology 2006;67:1189Y94



http://www.uab.edu/citherapy/training-manuals-a-publications



http://apps.who.int/classifications/icfbrowser/

http://apps.who.int/classifications/icfbrowser/

pediatric arm function test

Documents