Clinical Rehabilitation 2003; 17: 418-430

Wheelchair skills tests: a systematic reviewOlga JE Kilkens. Marcel WM Post iRv, Institute for Rehabilitation Research, Hoensbroek, Annet J Dallmeijer Institutefor Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit,Amsterdam, Henk AM Seelen iRv, Institute for Rehabilitation Research, Hoensbroek and Rehabilitation FoundationLimburg, Hoensbroek and Lucas HV van der Woude Institute for Fundamental and Clinical Human MovementSciences, Faculty of Human Movement Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Received 11th May 2002; returned for revisions 11th November 2002; revised manuscript accepted 14th December2002.

Objective: To describe and compare the content, feasibility, outcomeparameters, and clinimetric properties of the manual wheelchair skills testsreported in the literature.Design: A systematic literature search was conducted in MEDLlNE, EMBASE,PsychlNFO and Current Contents. Tests were selected if they wereobservational tests, designed for subjects using hand-rim wheelchairs andwere intended to assess wheelchair skill performance at the activity level.Results: The search resulted in 34 papers, in which 24 different wheelchairskills tests were described. The skill most frequently included was wheelchairpropulsion, consecutively followed by transferring, negotiating kerbs,ascending slopes, traversing tracks, sprinting and performing a wheelie. Thethree most frequently used outcome parameters were task performance time,independency of task performance, and physical strain during skillperformance. Sensitivity to change was evaluated in three tests, validity in10 tests, and reliability in nine tests.Conclusions: Many tests are applied to measure wheelchair skill performanceusing different tasks and outcome measures. This makes it difficult tocompare study results. Consensus among researchers as to which skills mustbe included as well as to standardization of the use of measurement

instruments will reduce this problem and will additionally lead to a betterinsight in the quality of tests.

Introduction

The achievement of independent mobility is vitalin the rehabilitation of physically disabled indi-viduals. When ambulation is impaired, a hand-rim wheelchair provides a relatively fast and

Address for correspondence: Olga Kilkens, Institute forRehabilitation Research, PO Box 192,6430 AD Hoensbroek,The Netherlands. e-mail: olga.kilkens@irv.nl

@ Arnold 2003

effective means of mobility for people with lowerlimb disabilities. A hand-rim wheelchair can pro-vide the necessary access to social, vocational andrecreational activities that are conditional to aproductive and rewarding life. To function inde-pendently, people who use manual wheelchairsfor mobility must possess a variety of skills. Theability to propel their wheelchairs over even sur-faces brings the freedom to move about within awheelchair-accessible environment. Independent

10.1 191iD2692 I 5503cr6330a

mobility within a greater variety of environmentsrequires obstacle negotiation skills. These skillscan make the difference between dependenceand independence in daily life.1.2

Assessment of wheelchair skills can provideuseful information concerning a person's currentwheelchair skill performance. In clinical situa-tions, wheelchair skills tests can help to definerehabilitation goals concerning mobility, and canalso be used to evaluate the progression maderegarding wheelchair mobility during rehabilita-tion. In research settings, measurement of wheel-chair skills can be used to study the effect of anintervention aimed at wheelchair mobility or tostudy the relation between wheelchair skills and,for example, level of activity and/or participation.

At present there is no systematic overview ofwheelchair skills tests available in the literature.It is therefore difficult to decide which test ismost suitable in research or in clinical practice.

The objective of this review is to systematicallydocument and describe the content, the targetpopulation, the study group, the test feasibility,the outcome parameters and the clinimetric prop-erties of those hand-rim wheelchair skills teststhat are currently reported in the literature. Suchan overview may make it easier to choose themost suitable test to assess wheelchair skills inboth clinical and research settings.

Methods

Search strategyTo locate wheelchair skills tests, a computer-

ized literature search of MEDLINE (1966-2001),EMBASE (1989-2001), Psych INFO (1967-2001)and Current Contents (1998-2001) was con-ducted. The keywords used were: mobility andwheelchair combined with skill, task, measure-ment, test. ADL, functional, instrument, perfor-mance, clinimetrics, psychometrics, pathology,behaviour, activity, disability and assessment.The search strategy is described in the Appendix.In addition, the references given in relevant pub-lications were further examined. Only studieswritten in English that were published in scien-tific journals were taken into consideration.

Wheelchair skills tests 419

Selection criteriaA test was selected if it was an observational

test,3 if it was constructed for subjects using hand-rim wheelchairs and when it intended to assesswheelchair-assisted mobility skills at the activitylevel as described in terms of the InternationalClassificiation of Functioning, Disability andHealth (ICF).4 In the ICF, mobility is defined as:'moving by changing body position or location orby transferring from one place to another'. Con-sequently, this review focuses on tests that aim toassess the ability to propel and manoeuvre awheelchair under standardized and/or simulatedconditions of daily living. Tests aimed at mea-suring physical capacity were not selected.

The first author performed the selection of thetests, by reading the abstracts of all the initiallyidentified articles. When necessary the full articlewas obtained and studied. In case of doubt onselection of a test, the other authors were con-sulted.

Assessment of selected testsThe wheelchair skills tests were systematically

described and compared with respect to the fol-lowing aspects:. Content: the skills included in the test.. Target population: the diagnostic groups for

which the test was developed.. Population at study: the diagnostic groups inwhich the test was used or studied.. Feasibility: the amount of time and equipmentneeded to perform the test.

. Test outcomes: the outcome parameters usedto reflect wheelchair skill performance and thecomplexity and interpretation of the scoringmethod.. Clinimetric properties: sensitivity to change,validity and reliability of the test.

Results

Selection of testsThe selection process produced 34 papers in

which 24 different wheelchair skills tests weredescribed.5-38Table 1 provides an overview of theselected tests, arranged alphabetically, accordingto the name of the first author of the paper inwhich the test was mentioned. Of the 24 tests

Table 1 General overview of the selected wheelchair skills tests

+-N0

ac.....

~~~(';;;::'"CD....

~

AuthorName of test

Outcomes

Agre"Bolins

Capodagll07Taricc033.34VFM

DallmeijerB.9

DuffililO

DunkerleyllDuran 12

Findley13Gans'4HaleylS.17TAMP

GouvierlbWebster36-38woeHarveyl8Hutllerl9

Janssen'O

Janssen"Janssenn.!ebsen23

2 Patients with72, 87, 55 stroke

20 No information9 Wheelchair

athletes37 No Information

44 No information37 No Information118 No information

Patients with stroke, healthy individuals

Patients with SCI

Wheelchair basketball players with poliomyelitis, SCI,and amputationPatients with SCI

Patients with SCIPatients with SCI

Patients with stroke, amputations, SCIneuropathy, hip fractures and healthy individuals

Wheelchair propulsion velocityTask performance timesPhysical strain (HRpeak)Subjective rating of performanceIndependency of task performance

Task performance timesPhysical strain (%HRR)Ability to perform tasksTask performance timesTask performance timesAbility to perform tasksWheelchair propulsion velocityTask performance timesIndependency of task performanceTechniqueMovement patternMovement control and co-ordinationNumber of collision errors

Independency of task performanceTask performance timesDistance covered

Task performance timesPhysical strain (HRpeak, HRmean)Physical strain (%HRR)Physical strain (%HRR)Task performance times

N Target population Study population

33 No information Patients with spina bifida4 No information Patients with SCI

8 Patients with SCI Patients with SCI47,94

20, 19 No information Patients with SCI

17 No information Patients with SCI11 No information Patients with SCI13 No information Patients with SCI

40 No information Patients with spina bifida, healthy individuals40 No information Patients with muscoloskeletal and206,206 neuromuscular disorders

Mlzukami28Panikoff29Schnelle3O.31Simmons32

1098097,7665

Vanlandewijck3o 46

No Information

No information

No information

No information

No informationNo informationNo information

Wheelchairathletes

Patients with sri, amputations and healthy individuals

Patients with stroke, amputations, srI.musculoskeletal and neuromuscular disordersPatients with sri

Patients with stroke, vascular diseasesand healthy Individuals

Patients with sriPatients with head injuryNursing home residents

Wheelchair basketball players with spina bifida, SrI.spastic diplegia, polio,amputations and healthyindividuals

Task performance timesPerceived task difficultySafety of skill performance% subjects able to learn skillAbility to perform tasks

Distance covered

Maximum angle of inclination of slopesTask performance timeDistance covered

Physical strain (HRpeak, physiologicalcost)

Perceived exertion during testperformance

Independency of task performanceIndependency of task performanceMaximum propulsion endurance timeIndependency of task performanceMaximal wheeling distanceWheelchair propulsion velocityTask performance timeDistance covered

Wheelchair propulsion velocity

sri, Spinal cord injury; VFM, Valutazione Funzionale Mielolesie; TAMP, Tufts Assessment of Motor Performance; WOC, Wheelchair Obstacle Course;WST, Wheelchair Skills Test; HRpeak, peak heart rate reached during task performance; %HRR, percentage heart rate reserve; HRmean, mean heartrate reached during task performance.

~~~-<'"1::r-;:,~.""~E;:~""r;,

.j:..N......

Klrby2" 42

Klrby2" 24WSTLehmann26 12

Mattison27 26

422 OJ E Kilkens et al.

found, seven were presente8 as measure-ment instruments and were extensively describedin terms of development, content anduse.7.14.18.23.25.35.36In all other papers the aim wasto evaluate an intervention or to detect differ-ences between groups. These tests were onlybriefly described in the Methods section of thearticle. Only four tests had been given aname: the Valutazione Funzionale Mielolesie(VFM),7 Tufts Assessment of Motor Perfor-mance (TAMP),14 the Wheelchair Skills Test(WST)25 and the Wheelchair Obstacle Course(WOC).36

Assessment of selected testsContent of tests

Table 2 displays the types of wheelchair skillsincluded in the different tests. Wheelchairpropulsion is the most frequently included skill(in 14 tests). It is assessed in different ways: a setperiod of time,6.19a fixed distance5.6.14.18.25.36or thelongest distance possible.3O.35Following wheel-chair propulsion, transfer from and to the wheel-chair is the most commonly included skill (in 11tests). Most tests require the performance ofseveral different transfers.7.14.18.20-23.28.29The nego-tiation of kerbs, ascending slopes and traversingtracks are third in line of most frequently usedskills (each in 10 tests). The height of the kerbsused ranged from 0.025 to 0.15 m. Two tests7.12require buth ascending and descending of thekerb. All other tests only assess the ascending ofthe kerb. In all but three tests,7.10.14the slopesused are defined in terms of inclination andlength, inclinations ranging from 1 to 11 degrees,length ranging from 3.05 to 21 m. Some examplesof tracks used are: slalom,6,19.25figure of eightlland obstacle course.12.36In six tests a sprint isincluded. Nearly all tests use a sprint over a fixeddistance (length ranging from 6.5 to 30 m).Although performing a wheelie is an importantskill in achieving wheelchair mobility, this skill isonly included in four tests.

Eleven tests include, in addition to the skillsalready mentioned, other specific wheelchairskills, e.g., managing brakes, negotiating doorsand loading the wheelchair into a car. Fifteentests consist entirely of the performance ofwheelchair skills. In eight tests wheelchair skillsare a part of a broader measure of ADL skills:

these tests encompass other ADL tasks such aseating, bed mobility skills and washinghands.7.8.14.18.21.2328.29

Target population and population at studyAlthough only four tests were designed for a

specific target population, 16 tests have only beenused in study groups with one specific diagnosis,most often spinal cord injury. Four tests wereused for subjects with varying medical conditions(Table 1).

FeasibilityOn the one hand, tests should include enough

elements to obtain an in-depth insight into wheel-chair skill performance; on the other hand, testshave to be efficient and as short as possible. Thecompletion time was mentioned for only six tests.The VFM,7 the TAMp14 and the test of Jebsen etal.23 take up to 1 hour to complete. However,these tests contain other ADL tasks as well asspecific wheelchair skills. The performance ofHarvey's testl8 requires approximately 15 min-utes, the time needed to complete the WST25 is30 minutes, and the mean test duration of thewheelchair basketball field test of Vanlandewijcket al.35is 1 hour and 22 minutes.

Ideally, tests should not require much space orspecial equipment. In most studies, the materialsneeded for test performance are not addressed.In their paper, Jebsen et al.23dedicated a sectionto test equipment (a hospital bed, standardizedwheelchair and straight chair). Harvey et al.18stated that no special equipment is required toperform their test. To assess physical strain dur-ing wheelchair skill performance a heart ratemonitor is required. Twelve studies provideinformation on the wheelchairs used during testperformance.6.8,1l-13.20-26Three studies used stan-dardized wheelchairsp.24.26 In eight studies, sub-jects used their daily use wheelchairs.8.11-13.20-2225Bolin et al.6 aimed to improve the individual fitof the wheelchair in their subjects. The subjectsperformed a wheelchair skills test twice: first intheir daily use wheelchair and later in an adaptedor new wheelchair.

The outcome measures of the different testsare displayed in Table 1. The most common out-come measure is task performance time. Inde-pendence in wheelchair skill I performance is

assessed by taking into account the use of assis-tive devices or the amount of help needed fromanother person. Six tests measure physical strainduring skill performance, four tests evaluate dis-tance covered during wheelchair propulsion, fourtests rate the velocity of wheelchair propulsion,and three tests assess subjective ratings regardingskill performance.

A test should preferably have an uncompli-cated scoring system that is convenient to use andthat can be analysed easily. The scoring of theTAMp14 is extremely complex: one hundred andthirteen skills have to be rated on six measure-ment dimensions and rating this test requiresextensive training.

Clinimetric propertiesNine tests6.10.11,13.22.24.26.28.30were not evaluated

on any of the clinimetric properties. Three testsprovide information on sensitivity to change.7.8,12Only two tests, the VFM7 and the WST,25 wereexplicitly subjected to a validation study. Foreight other tests5.12.19.2L.24.2628.30.36information onvalidity could be retrieved from the articles. Thevalidity of these tests, however, was not explicitly

evaluated. Information concerning reliability wasgiven for 10 tests.7.14,15,18-20,23.25.27.35Only fivetestsI9.23.25.27.36provided data on both reliabilityand validity. Table 3 displays the 11 tests fromwhich the sensitivity to change and/or the valid-ity have been evaluated. In Table 4, the 10 teststhat have been assessed on the topic of reliabil-ity are shown.

From Table 4 it can be seen that all availabletest-retest and inter-rater reliability figures aresatisfactory up to excellent; the data on validityare less unequivocal (Table 3).

Discussion

A literature search resulted in the selection of 24different wheelchair skills tests. This collectionmay be incomplete, since only English-writtenstudies, published in scientific journals weretaken into account. However, we feel that wehave provided a critical and useful overview oftests in which wheelchair skills are assessed.

Wheelchair skills tests 423

Table 2 Content of wheelchair skills tests

Author Propulsion Sprint Kerb Slope Wheelie Transfer Track Other wheelchair skills

Agre5 + +Bolin6 + + + +Capodaglio'; Taricco32.33 + + + + + +Dallmeijef.9 + + +Duffilll° + + + + +Dunkerley'l + +Duran'2 + + + + +Findley13 + +Gansl4; Haleyl6,17 + + + +Gouvier'5; Webster36-38 +Harvey18 + + + +Hutzler19 + + +Janssen2o + +Janssen21 + + + +Janssen22 + +JebseoZ3 + + +Kirby24 +Kirby25 + + + + + + +lehmann26 + +Mattison27 +Mizukamj28 + + +Panikoff29 + + + + +Schnelle3o,31;Simmons32 +Vanlandewijc125 + + +

less common wheelchair skills are noted in the column 'other wheelchair skills',

Table3 Sensitivityandvalidityof the selectedwheelchairskillstests

+>-N+>-

~~~~<':>:::'"('t>....

~

Agre5 No information

ConstructvalidityAuthor Sensitivityto change Contentvalidity

Capodaglio7Taricco3!.34

Two groupswere tested beforeandaftera6-weekrehabilitationprogramme:conventionalor enhanced.No significantdifferencesbetween the groups7Subjectswere tested at the start andtheend of their rehablitationperiod.Scoreshadsignificantlyimprovedfor subjectswithquadriplegiaand high-levelparaplegia.Therewere no changesin the scoresofsubjectswith low-levelparaplegia3!Sublectswere tested at the start andat theend of their rehabiliationperiod.Therewereno changesin physicalstrainduringtheslopeascendandthe transfers.Theperformancetime of the slopeascendsignificantlydecreased.Theperformancetime of the transferdid not change"Subjectswere tested beforeand aftera 16-week exerCiseprogramme.Taskperformancetimes significantlydecreased.The abilitytonegotiatekerbsimprovedNo information

Dallmeijer"."

Duran 12

Gouvler'.Webster""""

Hutzler'. No information

Janssen" No information

No information

Thecontentof the VFM was comparedto thecontentof the FIM. BarthelIndexandaiF. TheVFM includesall basicADLdomainsandhasthe largestnumberof tasks in the domainstransfersandwheelchairuse which areparticularlyrelevantfor wheelchair-dependentpeople3!

No information

No information

No information

No information

No information

Failedto find a relationshipbetweenmotor functionand propulsionspeedNo information

No information

Mean FIM scoreswere associatedwith an improvedability to negotiatekerbs,andinverselyassociatedwithtask performancetimes

Right-sidedstroke patientswith neglecthit significantlymore objectsin !eft- thanin right-space,andmadesignificantlymoreerrorsthanstroke patientswithoutneglectand healthycontrolsubjects36.37Inversecorrelationbetweensplint performancetimeand the subjects'wheelchairbasketballclassification

Ir"., ,...,=-0.641. No relationbetweenbodyweight andtask performancetimesA lower lesionlevelwas associatedwith a lower levelof physicalstraindurrngwheelchairskill performance.No relatIonbetweenlesion-completenessandphysicalstrain duringskill performance.Physicalstrainduringskill performancewas inverselyrelatedto strength,sprint power,peakoxygenuptake,

and maximumpower output(rf"""'M=-0.41to -0.731

Jebsen23 No information

Klrby25 No inform atlon

Mattlson27 No information

Panikoff29 No information

FIM, FunctionalIndependenceMeasure;QIF,QuadriplegiaIndexof Function,

No information

OccupatIOnaltherapistsevaluatedthe contentof the WST,andunanimouslyapprovedwith 30of the 33 skills

No informatlon

No inform ation

Healthy people performed the tasks concerningwheelchair mobility faster than patients with mobilityproblemsOccupational therapists rated whether patients'wheelchall skills had Improved, not changed orworsened between two test tllals, They rated that 13subjects had Improved, 8 had not changed, and nonehad worsened, Accordingly, the mean improvement Inthe total score was larger for the 'improved subjects'than for the 'unchanged subjects', OT's also completeda VAS reflecting the subJects' wheelchair skillperformance, Mean improvement for the 'improvedsubJ8cts' was significantly higher than for the'unchanged subjects',Test scores significantly related to age, and wheelchairexperience, No relation was found between test scoresand dragnosisThere was no relation between physical strain duringskill performance and age or diagnosIs, Physicalstrain was significantly related to the perceived exertionduring test performance Ir =0,841A significant relation betw=the ability to performtransfers and length of coma

~"""";:,~s::,:::;.'"~;::,;<::;-

~c::;

.j;:.NV\

426 OJE Kilkens et al.

Clinical messages

. A large number of wheelchair skills testshave been described in the literature.. It is recommended to use only those teststhat have been tested and shown to be ade-quate in terms of responsiveness, validityand reliability.. Future research should focus on further val-idation of existing tests instead of develop-ing more tests.

Content of testsThere is limited consensus as to the content of

wheelchair skills tests (Table 2). Even skills fre-quently used in tests (wheelchair propulsion,transfer, kerb, slope) show a large variation in,for instance, driving distance, objects to transferto, height of the kerbs and angle of inclination ofslopes. The number of skills included in wheel-chair skills tests also shows a large variation,ranging from one36 to 113.14 More research isneeded to identify a limited number of skills thattogether best reflect wheelchair skill performancein people who dependent on a manual wheelchairfor their mobility.

Target population and population at studyAlthough just four tests were designed for a

specific target group, 16 tests were only used ina specific group (Table 1). The latter tests mightalso be capable of assessing wheelchair skills insubjects with other characteristics, but furtherresearch on validation and reliability in othersubject groups is necessary to test this expecta-tion.

Test outcomesMany different outcome parameters are used

in the selected tests (Table 1). The choice for aparticular outcome measure depends on theobjectives of the study. Tests can be used todetermine the feasibility of manual wheelchairpropulsion, to measure the level of independencein wheelchair ADL, or to evaluate the effects ofinterventions. Outcome measurements in thecategories time, distance and physical strain arevery useful to provide information on thepracticability of manual wheelchair mobility in

daily life. When a subject needs, for example, fiveminutes to propel his or her wheelchair over adistance of 50 metres, an electric wheelchair maybe a more suitable means of mobility. If the goalof a study is to describe the level of independentmobility, a scale of independence in performingwheelchair tasks is an obvious outcome measure.

The level of independence in performing well-chosen wheelchair skills is expected to be directlyrelated to independent mobility in daily life.People who cannot perform wheelchair skillsindependently will not achieve independentmobility in all environmental circumstances. Forthe assessment of (changes in) wheelchair skillperformance in completely independent individ-uals, outcome measures such as time, distanceand physical strain should be applied. This is alsoshown by the results of Taricco et al.34who mea-sured wheelchair skill performance using a scaleof independence. They showed good sensitivityto change in subjects with high-level spinal cordinjury, but no sensitivity to change in subjectswith low-level spinal cord injury.

Other outcome measures are relevant, but notso easy to interpret. The test of Dallmeijer et al.8.9evaluates both physical strain during wheelchairskill performance and performance time of eachskill. These two parameters are, however, inter-dependent. A decrease in performance time,reflecting better test performance, may result ina higher level of physical strain, indicating worsetest performance. This interdependency mayobfuscate the interpretation of test results.

Further, wheelchair skill performance relies onboth technique and physical capacity. Repeatedmeasurements can, for instance, show that maxi-mal wheeling endurance time has increased overa certain period, which may be the result of anincrease in physical capacity, an improved tech-nique resulting in higher mechanical efficiency ofwheelchair propulsion, or a result of both. For acorrect interpretation of changed outcomes inlongitudinal studies, the performance of a wheel-chair skills test is best combined with an exercisetest that provides information about physicalcapacity.

The WST25leads to one overall score of wheel-chair skill performance, expressed as the sum ofthe scores obtained on each skill. Such a totalscore might be very useful for research purposes,

rWl

Author

Reliability of the selected wheelchair skills tests

Inter-rater reliability

Table 4

Test-retest reliability Intra-rater reliability

Capodagli07Taricc033.34

Gans14

Haley",17

Gouvier15Webster3"38

Harvey'8

Hutzler19

Janssen"

Jebsen23

Kirby25

Mattison"

Vanlandewijck35

No Information

No information

No information

No information

Excellent test-retest reliability for taskperformance time and propelled

distance (r,pe"m,,=0.87-D.99) on twotest performancesSubjects performed three test trials:two on the same day (trial 1 and 2) andone a week later (trial 3). Correlations

for heart rate were good for trial 1versus 2. and trial 1 versus 3 (1 vs 2:r =0.79-0.97. ICC=0.73-D.97. 1 vs

3e'~:::,,0, =0.69-0.95. ICC=0.52-D.92).Good correlations for the performancetime of the kerb ascend for trial 1

versus 2 !rpe",o,=0.88. ICC=0.82). Lowcorrelations for trial 1 versus 3

(rpe"o, =0.31. ICC=0.20)Performance times correlations wereexcellent (r =0.85-D.99)Good correr~t::;'~ between the scoreson two different test occasions

(r,pe"m,,=0.65)Subjects performed two test trials. Goodcorrelations for distance traveled.physical strain during wheelchairpropulsion. and propulsion velocity

(~""'0,=0.96, 0.84. and 0.70)Excellent correlations for all tasks

(r,pe"meo=0.8O-D.97). only one beinggood (r,pe"m,,=0.65)

The inter-rater reliability of the VFM wasdemonstrated In previous studies. These studieswere only published in Italian and were thereforenot retrieved

Subjects' test performances were videotaped, andindependently evaluated by three raters. Regardingwheelchair skills. inter-rater reliability was excellent(kappa = 0.65-D.83)13Two raters scored the number and type oferrors two subjects made during test performance.Their agreement on the occurrence of errors and thetype of error made was successively 95% and 83%14.Two raters scored the tests. They agreed100% on the occurrence of errors, 80-90% on

the occurrence of direct hits (roe","0=0.90-0.97). and85-90% on the occurrence of sideswipes errors

(rpe"", =O.92-D.95)35Two raters scored the test performances resulting inhigh inter-rater reliability (kappa=O.82-D.98)No information

No information

No information

Two raters independently scored the same videotapesof test performances, resulting in an excellent

correlation coefficient (r,pe"m" =0.95)No Information

No information

No information

No mformation

No information

No information

No mformation

No information

No information ~t1>t1>

1=;";::s-~:::;.'";>;-;::.;c;;-

~'"~

One rater scored the same videotapes of subjectstest performances twice. resulting in an excellent

correlation coefficient (r,p",m" =0.96)No information

No information

+:-N-.:t

428 OlE KiIkens et al.

but can only lead to valid information if all testitems measure the same phenomenon. Theauthors did not assess this.

Clinimetric propertiesClinimetric properties of nine tests were not

described at all. Only two tests: the VFM7 andthe WST25 have been extensively validated. TheWST is the only test that has been adequatelytested on both validity and reliability. Moreresearch is needed to assess the clinimetric qual-ities of the other tests described in the currentreview before these tests can be recommendedfor use. Performance time and physical strain areoutcome measures that can be measured objec-tively. Ordinal scales of dependence, frequentlyused in wheelchair skills tests, are subject tointerpretation. Raters need to assess the amountof help, often expressed in a few number of cat-egories of assistance needed. Therefore, theraters should receive appropriate training. Thegood inter-rater reliability figures of tests usingratings of independence are promising, but thesefigures are, in part, obtained in very small studygroups. Although also rarely investigated,test-retest analyses of time, distance, velocity andphysical strain revealed satisfying results.

The measurement of wheelchair skills will, atleast in part, support validity of the tests due tothe close resemblance with daily life activities.However thc choice of tested tasks, outcomemeasures and the applicability in different sub-ject groups may influence validity. One aspect ofvalidity that is often ignored is the influence ofwheelchair configuration on wheelchair skill per-formance outcomes. Subjects will perform best ina wheelchair that is optimally adjusted to theirpersonal characteristics. To ensure that variationsin wheelchair skill performance were not due tochanges in wheelchair configuration, some testswere performed in standardized wheel-chairs.23.24.26However, most tests were executedin daily use wheelchairs.8,12.2o.22.25This may haveresulted in subjects using different wheelchairson different test occasions, which may haveaffected sensitivity to change and test-retest reli-ability or may bias comparisons between subjectshaving wheelchairs of different quality. Use ofthe daily use wheelchair may, however, improvethe validity of the test. Subjects are not troubled

by an unfamiliar wheelchair, and their test resultswill be more representative for their wheelchairskill performance in daily life. Therefore a care-ful choice for, or against standardization ofwheelchair configuration has to be made, depen-dent on the purpose and the design of the study.

In conclusion, this review shows that there is,as yet, no standard test to measure wheelchairskill performance. Only seven out of the 24 testsfound were extensively described in terms ofdevelopment, content and use7.14.18.23.25.35.36andonly two tests have been extensively validated.1.25

In addition, most tests have only been used inone study. Without further research on validityand reliability, these tests should be used withcaution. The use of many different tests makes itdifficult, if not impossible, to compare studyresults. Standardization of the skills tested andthe use of measurement instruments are neededto enable comparisons between studies and togive a better insight in the quality of the testsused.

Future research could best concentrate on fur-ther validation of existing tests instead of devel-oping even more tests. The selection of the bestand most relevant items of these tests and com-bining elements of various tests might eventuallylead to a superior test. However, it might not bepossible to compose the ideal test for all patientgroups and purposes. A distinction between aclinical instrument (containing all relevant itemsfor assessment and evaluation of individualtreatment) and a research instrument (containinga selection of items of varying difficulty) mightbe useful.

AcknowledgementsThe study was supported by the Dutch Health

Research and Development Council, ZON-MW,Rehabilitation program, grant no. 1435.0003.

References

1 Somers M. Spinal cord injury, functional rehabilita-tion. Connecticut: Appleton & Lange, 1992.

2 Britell CWoWheelchair prescription. In: LehmannJF, Kottke FJ eds. Krusen's handbook of physicalmedicine and rehabilitation, fourth edition. Philadel-phia: WB Saunders, 1990:548-63.

3 Bussmann JBJ, Starn HJ. Techniques for measure-ment and assessment of mobility in rehabilitation: a

theoretical approach. Cfin Rehabil 1998;12: 455-64.4 WHO. International Classification of Functioning,

Disability and Health. IClDH-2 final draft. Geneva:WHO, 200l.

5 Agre JC, Findley TW, McNally MC et al. Physicalactivity capacity in children with myelomeningocele.Arch Phys Med Rehabil1987: 68: 372-77.

6 Bolin 1, Bodin P, Kreuter M. Sitting position -posture and performance in C5-C6 tetraplegia.Spinal Cord 2000; 38: 425-34.

7 Capodaglio P, Grilli C, Bazzini G. Tolerableexercise intensity in the early rehabilitation ofparaplegic patients. A preliminary study. SpinalCord 1996;34: 684-90.

8 Dallmeijer AJ, van der Woude LHV, Hollander AP,Angenot ELD. Physical performance in persons withspinal cord injuries after discharge from rehabilita-tion. Med Sci Sports Exerc 1999;31: 1111-17.

9 Dallmeijer AJ, van der Woude LHV, Hollander AP,van As HHJ. Physical performance during rehabili-tation in persons with spinal cord injuries. Med SciSports Exerc 1999;31: 1330-35.

10 Duffill J, Buckley J, Lang D, Neil-Dwyer G,McGinn F, Wade D. Prospective study of omentaltransposition in patients with chronic spinal injury. JNeurol Neurosurg Psychiatry 2001; 71: 73-80.

11 Dunkerley AL, Ashburn A, Stack EL. Deltoidtriceps transfer and functional independence ofpeople with tetraplegia. Spinal Cord 2000; 38:435-4l.

12 Duran FS, Lugo L, Ramirez L, Lic EE. Effects ofan exercise program on the rehabilitation of patientswith spinal cord injury. Arch Phys Med Rehabil2001; 82: 1349-54.

13 Findley TW, Agre Je. Ambulation in the adolescentwith spina bifida. II. Oxygen cost of mobility. ArchPhys Med Rehabil 1988;69: 855-6l.

14 Gans 8M, Haley SM, Hallenborg SC, Mann N,Inacio CA, Faas RM. Description and interobserverreliability of the Tufts Assessment of MotorPerformance. Am J Phys Med Rehabil1988; 67:202-10.

15 Gouvier WD, Cottam G, Webster JS, Beissel GF,Wofford JD. Behavioral interventions with strokepatients for improving wheelchair navigation. Int JClin NeuropsychoI1984; 6: 186-90.

16 Haley SM, Ludlow LH, Gans BM, Faas RM, InacioCA. Tufts assessment of motor performance: anempirical approach to identifying motor performancecategories. Arch Phys Med Rehabill991; 72: 359-66.

17 Haley SM, Ludlow LH. Applicability of thehierarchical scales of the Tufts Assessment of MotorPerformance for school-aged children and adultswith disabilities. Phys Ther 1992;72: 191-202;discussion 202-206.

18 Harvey LA, Batty J, Fahey A. Reliability of a toolfor assessing mobility in wheelchair-dependentparaplegics. Spinal Cord 1998;36: 427-31.

Wheelchair skills tests 429

19 Hutzler Y. Physical performance of elite wheelchairbasketball players in armcranking ergometry and inselected wheeling tasks. Paraplegia 1993;31: 255-61.

20 Janssen TWJ, Van Oers CAJM, Van der WoudeLHV, Hollander AP. Reliability of heart rateresponses to non-steady-state activities of daily livingin men with spinal cord injuries. Scand J RehabilMed 1994;26: 71-78.

21 Janssen TWJ, van Oers CAJM, Veeger HEJ,Hollander AP, van der Woude LHV, Rozendal RH.Relationship between physical strain during stan-dardised ADL tasks and physical capacity in menwith spinal cord injuries. Paraplegia 1994; 32:844-59.

22 Janssen TWJ, van Oers CAJM, Rozendaal EP,Willemsen EM, Hollander AP, van der WoudeLHV. Changes in physical strain and physicalcapacity in men with spinal cord injuries. Med SriSports Exerc 1996;28: 551-59.

23 Jebsen RH, Trieschmann RB, Mikulic MA, HartleyRB, McMillan JA, Snook ME. Measurement of timein a standardized test of patient mobility. Arch PhysMed Rehabil1970; 51: 170-75.

24 Kirby RL, Lugar JA, Breckenridge C. New wheelieaid for wheelchairs: controlled trial of safety andefficacy.Arch Phys Med Rehabil 2001; 82: 380-90.

25 Kirby RL, Swuste J, Dupuis DJ, MacLeod DA,Monroe R. The Wheelchair Skills Test: a pilot studyof a new outcome measure. Arch Phys Med Rehabil2002: 83: 1()-18.

26 Lehmann JF, Warren CG, Halar E, Stonebridge JB,DeLateur BJ. Wheelchair propulsion in thequadriplegic patient. Arch Phys Med Rehabill974;55: 183-86.

27 Mattison PG, Hunter J, Spence S. Development of arealistic method to assess wheelchair propulsion bydisabled people. Int J Rehabil Res 1989; 12: 137-45.

28 Mizukami M, Kawai N, Iwasaki Y et al. Relationshipbetween functional levels and movement intetraplegic patients. A retrospective study.Paraplegia 1995;33: 189-94.

29 Panikoff LB. Recovery trends of functional skills inthe head-injured adult. Am J Occup Ther 1983;37:735-43.

30 Schnelle JF, MacRae PG, Ouslander JG, SimmonsSF, Nitta M. Functional Incidental Training, mobilityperformance, and incontinence care with nursinghome residents. J Am Geriatr Sac 1995;43: 1356-62.

31 Schnelle JF, MacRae PG, Giacobassi K, MacRaeHSH, Simmons SF, Ouslander JG. Exercise withphysically restrained nursing home residents:maximizing benefits of restraint reduction. J AmGeriatr Sac 1996:44: 507-12.

32 Simmons SF, Schnelle JF, MacRae PG, OuslanderJG. Wheelchairs as mobility restraints: predictors ofwheelchair activity in nonambulatory nursing homeresidents. J Am Geriatr Soc 1995;43: 384-88.

33 Taricco M, Colombo C, Adone R, Chiesa G, Di

430 OlE Kilkens et aL

Carlo S, Borsani M et al. The social and vocationaloutcome of spinal cord injury patients. Paraplegia1992;30: 214-19.

34 Taricco M, Apolone G, Colombo C, Filardo G,Telaro E, Liberati A. Functional status in patientswith spinal cord injury: a new standardizedmeasurement scale. Gruppo InterdisciplinareValutazione Interventi Riabilitativi. Arch Phys MedRehabil 2000; 81: 1173-80.

35 Vanlandewijck YC, Daly DJ, Theisen DM. Field testevaluation of aerobic, anaerobic, and wheelchairbasketball skill performances. Int J Sports Med 1999;20: 548-54.

36 Webster JS, Cottam G, Gouvier WD, Blanton P,Beissel GF, Wofford J. Wheelchair obstacle courseperformance in right cerebral vascular accidentvictims. J Clin Exp Neuropsychol1989; 11: 295-310.

37 Webster JS, Rapport LJ, Godlewski MC, AbadeePS. Effect of attentional bias to right space onwheelchair mobility. J Clin Exp NeuropsychoI1994;16: 129-37.

38 Webster JS, Roades LA, Morrill B et al. Rightwardorienting bias, wheelchair maneuvering, and fall risk.Arch Phys Med RehabiI1995; 76: 924-28.

Appendix - Search strategy

#1#2#3#4#5#6#7#8#9#10#11#12#13#14#15#16#17

mobility and wheelchair#1 and ski1l*# I and task*#1 and measurement*#1 and test*#1 and ADL#1 and functional*#1 and instrument*#1 and performan<:e# 1 and clinimetrics#1 and psychometrics#1 and pathology#1 and behaviour#1 and activit*#1 and disabilit*#1 and assessment#2 or #3 or #4 or #5 or #6 or #7 or #8 or #9or #10 or #11 or #12 or #13 or #14 or #15 or#16