The addenbrooke’s cognitive examination revised as a potential screening test for elderly drivers

(1)

ContentslistsavailableatSciVerseScienceDirect

Accident

Analysis

and

Prevention

j ou rna l h o me pag e :w w w . e l s e v i e r . c o m / l o c a t e / a a p

The

Addenbrooke’s

Cognitive

Examination

Revised

as

a

potential

screening

test

for

elderly

drivers

Inês

S. Ferreira

a,∗

_,

_Mário

_R.

_Simões

b

_,

_João

_Marôco

c

a_Centro_de_Investigac¸_ão_do_Núcleo_de_Estudos_e_Intervenc¸_ão_Cognitivo_{Comportamental}_(CINEICC),_Faculty_of_Psychology_and_Educational_Sciences,_University_of_Coimbra,_Coimbra,

Portugal

b_Centro_de_Investigac¸_ão_do_Núcleo_de_Estudos_e_Intervenc¸_ão_Cognitivo_{Comportamental}_(CINEICC),_{Psychological}_Assessment_Lab.,_Faculty_of_Psychology_and_Educational_Sciences,

UniversityofCoimbra,Coimbra,Portugal

c_Instituto_Superior_de_Psicologia_Aplicada_–_Instituto_{Universitário,}_Lisboa,_Portugal

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received31October2011

Receivedinrevisedform20March2012 Accepted29March2012

Keywords:

Addenbrooke’sCognitiveExamination Revised Fitnesstodrive Prediction On-roadassessment Olderdriver

a

b

s

t

r

a

c

t

Considerableresearchhasshownthatneuropsychologicaltestsarepredictiveofreal-worlddriving abil-ity.TheMini-MentalStateExamination(MMSE)isabriefcognitivetestthathasbeencommonlyused intheassessmentofolderdrivers.However,thistesthasinherentproblemsthatlimititsvalidityto evaluatecognitiveabilitiesrelatedtodrivingandtoscreenfordrivingimpairmentsinnon-demented people.Therefore,itisusefultotestnewscreeninginstrumentsthatmaypredictpotentialunsafedrivers whorequireanin-depthneuropsychologicalassessmentinaspecialisedcentre.Todate,theutilityof theAddenbrooke’sCognitiveExaminationRevised(ACE-R)asanindicatorofdrivingabilityhasnotbeen established.Inthecurrentstudy,fiftyolderdrivers(meanage=73.1years)whowerereferredfora psy-chologicalassessment,theprotocolofwhichincludedtheACE-R,underwentanon-roaddrivingtest. Usinglineardiscriminantanalyses,theresultshighlightedthehigherclassificationaccuracyofthe ACE-RcomparedtotheMMSEscore,particularlyfordetectingunsafedrivers.Measuresofvisuospatialand executivefunctions,whicharenotincorporatedintheMMSEscore,hadanincrementalvalueinthe predictionofdrivingability.Thisemergingbriefcognitivetestmaywarrantadditionalstudyforusein thefitnesstodriveassessmentofolderadults.

1. Introduction

Trafﬁcsafetyforolderdriversisagrowingpublichealthissue.

The demographics of aging and life expectancygive rise toan

increasingnumberof olderlicenseddriversontheroad,asthe

personal automobileconstitutes theprimarymodeof transport

fortheemerging elderlydrivers(OrganisationforEconomic

Co-operationandDevelopment,2001).Olderdrivershaveahighrateof

trafﬁcfatality,causedbothbycrashinvolvementandinjury

sever-ity(EuropeanRoadSafetyObservatory,2006).Cognitivedeclineis

moreprevalentwithagingandcertainmedicalconditions(suchas

dementia),andtherelationbetweencognitivedeclineandgreater

crashriskandunsafedrivingperformanceiswelldocumented(Carr

andOtt,2010).

Investigationshaveshownthatneuropsychologicaltests

mea-sureabilitiesthatarerelevantforeveryday/real-worldfunctioning,

∗ Correspondingauthorat:PsychologicalAssessmentLab.,FacultyofPsychology andEducationalSciences,UniversityofCoimbra,RuadoColégioNovo,Apartado 6153,3001-802Coimbra,Portugal.Tel.:+351936343774;fax:+351256084386. E-mailaddresses:[email protected],[email protected](I.S.Ferreira).

includingdrivinganautomobile(MorganandHeaton,2009).For

instance,twometa-analysesoftheliterature,oneofpersonswith

dementia (Reger et al., 2004) and one of older adults without

a dementia diagnosis(Mathias andLucas,2009), provided

sup-port for the utility of tests of visuospatial abilities and visual

attention, respectively, in predicting real-world driving ability.

Tests of executive functions have also shown promise in

pre-dicting on-road (Ott et al., 2008; Whelihan et al., 2005) and

crashinvolvement (Daigneault et al.,2002), althoughexecutive

functioninghasnotbeenexaminedasextensivelyasother

neu-rocognitivedomains.Nevertheless,neuropsychologicaltestscores

donotcorrelateentirelywithfunctionaloutcomes(Sadekandvan

Gorp,2010),andcurrentinvestigationsattempttoconcludethe

mosteffectiveassessmentmethodsfordetectingdrivingdeﬁcits

(MorganandHeaton,2009).

The Mini-Mental State Examination (MMSE; Folstein et al.,

1975)isacognitivescreeningtestthathasbeenwidelyusedas

a potential predictor of driving ability (cf., Reger et al., 2004),

althoughitwasnotdesignedforthispurpose.Theinvestigations

have produced conﬂicting results. A report from the

Ameri-canMedicalAssociation(Iversonetal.,2010)concludedthatan

MMSEscorelowerthan25pointscouldbeusefulinidentifying

(2)

patientsatincreasedriskofunsafedriving;however,dueto

con-ﬂictingdata,any correlationwithdriving performance remains

unclear.

Severalseriouslimitationshavebeenidentiﬁedin respectto

theMMSEasameansofscreeningfordrivingimpairments.These

limitationsinclude thepoorsensitivity for those scoringatthe

mid-range (Marcotte and Scott,2009) and theweaker

associa-tionwithdrivingmeasuresatthehigherendofthescale(score

greaterthan 26) (Fitten et al.,1995).In addition, theMMSE is

basedalmostentirelyonverbalcognitivefunctionattheexpense

ofnon-dominanthemisphereskillsandexecutivefunctions(Bak

andMioshi,2007),whicharethemostrelatedtovariousdriving

measures(cf.,Amicketal.,2007;Daigneaultetal.,2002;Lincoln

etal.,2010;Regeretal.,2004;Whelihanetal.,2005).Moreover,

theassessmentofattention(verbalmodality)isquitesuperﬁcial.

Infact,whenadditionalmeasuresofattention,visuospatialand

executivefunctionsareconsideredalongwithMMSEscores,the

predictionaccuracyofdrivingabilityincreases(Ottetal.,2005).

Duetotheselimitations,anewbriefcognitiveinstrumenttoscreen

for driving-related cognitive impairment in the elderlyis

war-ranted.Suchaninstrumentwouldbeparticularlyusefulforprimary

carephysicianswhoareresponsiblefordecisionsregardingtheir

patients’ﬁtnesstodrive.Asdrivingﬁtnessreliesonafunctional

evaluationofmultipledomains(cognitive,motor,perceptual,and

psychiatric),nosingletestissufﬁcienttodeterminewhetheran

individual is unsafe to drive (Rizzo, 2011). However, a

cogni-tivescreeningtestcouldbeachievedtoidentifypotentialunsafe

drivers(Wolfe and Clark,2012)who requirefurther

neuropsy-chological(and,ifavailable,on-road)assessmentataspecialised

centre.

Cognitiveimpairmentisa recogniseddrivingriskfactor that

couldberelatedtodeﬁcitsinattention,perception,memoryand

executivefunctions(RizzoandKellison,2010),speciﬁcskillsthat

areassessedbyarecentcognitivescreeningtest,theAddenbrooke’s

CognitiveExaminationRevised(ACE-R;Mioshietal.,2006).

Driv-ing impairments might be expected in various populations, as

normal aging, clinical neurologic disease (namely, Alzheimer’s

Disease,Parkinson’sDisease,MultipleSclerosis,Stroke)and

psy-chiatricconditions(namely,MajorDepression,Schizophrenia)that

affecta person’s cognitivefunctioning and may impactdriving

ability(MarcotteandScott,2009).ACE-Rresearchhasexamined

theseconditions(e.g.,Alexopoulosetal.,2010;Dudasetal.,2005;

Hamiltonet al., 2009; Kwak etal., 2010; Mathew et al., 2011; Pendleburyetal.,2012;Reyesetal.,2009).

Thegoalofthecurrentstudyistoexaminethepotentialvalue

oftheACE-Rforpredictingdrivingabilityinolderdrivers.There

areanumberofreasonstoanalysethisbriefcognitivetestin

rela-tiontoon-roaddrivingperformance.Firstly,itisamoreaccurate

meansof detectingearlycognitivedysfunctionthantheMMSE,

withanumberofaddeditemstoincreaseitssensitivityto

spe-ciﬁcdementingdiseases(e.g.,Alexopoulosetal.,2010;Mioshietal.,

2006;Simõesetal.,2011;Torralvaetal.,2011;Yoshidaetal.,2012)

thatmightadverselyaffectdrivingabilitiesandincreasecrashrisk

(Iversonetal.,2010).Secondly,itcomprisesanMMSEscoreand

sub-scoresinﬁvecognitivedomains,whichcouldprovide

qualita-tiveanddiscriminativedataregardingthepatient’sdriving-related

cognitiveproﬁle.Thirdly,theACE-Rmayprovidesome

discrimi-natingvaluefordrivingability,asitincludestasksthataresimilar

tothosepreviouslyassociatedwithdrivingmeasures(cf.,DeRaedt

andPonjaert-Kristoffersen,2001;Johanssonetal.,1996;Marottoli

etal.,1994).Finally,theACE-Risabrief,inexpensive,andeasily

administeredtool.Ifitisvalidatedtopredictcrashriskordriving

ability,itwouldbeappealingtoprimarycarehealthsettings.The

currentresearchexamineswhetherthisscreeningtoolfor

cogni-tiveimpairmentisusefulinpredictingdrivingimpairmentinolder

adults.

2. Methods

2.1. Participants

ThecurrentinvestigationwasconductedatthePsychological

LaboratoryofMobility andLandTransportsInstitute(MLTI),the

uniquestatecentrethatperformspsychologicalexaminationsof

drivingﬁtnessinPortugal.Theparticipants’primarycarephysician

referredthemduetodoubtsabouttheirdriving-relatedcognitive

abilitiesduringaﬁtnesstodriveassessment.Theinclusion

crite-riaofthepresentstudywereasfollows:aged65yearsorolder;

livingindependently in thecommunity;withouta requirement

forspeciﬁcadaptiveequipmentfordriving;auditoryand/or

lan-guagefunctioningcompatiblewithverbalcomprehensionand/or

expression(pre-requisiteforon-roadandneuropsychological

test-ing);visualacuityinthecentralﬁeldofview(far-sight,withregular

correction)withinlicensingrequirementsinPortugal;andatleast

tenyearsof drivingexperience,includingdrivingexperiencein

thelastmonth.Duringthedatacollectionperiod(November2009

toDecember2010),theMLTIexamined83driversaged65years

or more, ofwhom 23 didnot take partin theresearch

proto-col:16personsdidnotmeettheinclusioncriteria(eightbecause

ofthepresenceofmotordeﬁcitsandneedsforspeciﬁcadaptive

equipmentfordriving,onedidnotfulﬁlthevisualacuity

licens-ingrequirements,two personwereaphasicwithseveredeﬁcits

inverbalcommunication),sixwerenotscheduledinthedaysof

theinvestigation,fourwerenot assessedduetotheabsenceof

thedrivingexaminerandtworefusedtoparticipate.Ofatotalof

60driverswhocompletedtheresearchprotocol,50participants

aged65–88years(M=73.1,SD=7.0)wereselectedtoobtainastudy

samplewithequallysizedgroupsdeﬁnedbythedependent

vari-able(safe/unsafe)andequallydistributedbythereasonforreferral

(aging,physical,neurologicalorpsychiatric).Thissampling

strat-egy wasdesigned tominimise thepotential negativeinﬂuence

ofunequallysizedgroupsonthestatisticalanalysis(discriminant

functionandclassiﬁcationcases)andtobalancetherangeof

rea-sonsphysiciansrequestedafurtherpsychologicalassessmentin

a specialisedcentre.The participantswerereferred for the

fol-lowingreasons:aging(n=24);physicaldisorders(n=4),including

rheumatoid arthritis(n=2),prosthesisonboth feet(n=1),

con-genitalmyopathy(n=1);neurologicaldisorders(n=12),including

Parkinsondisease(n=6),dementiaofAlzheimertype(n=2),stroke

(n=2),cerebralangioma(n=1)andmultiplesclerosis(n=1);and

psychiatricdisorders(n=10),includingdepression(n=10),remote

historyofalcoholabuse(n=3),residualschizophrenia(n=1)and

bipolar disorder (n=1). Participants had a mean education of

6.18years(SD=4.3;min=4;max=20).Astotheeducationallevel,

64%completedprimaryschool(fourthgrade),22%middleschool

(ninthgrade),2%highschool(twelfthgrade)and12%were

univer-sitygraduates.Thenumberofschoolyearscompletedbytheoverall

Portuguesepopulationover64yearsisasfollows:36.2%are

illit-erate,46.4%completeprimaryeducation,9.4%middleschool,3.0%

highschool,and5.0%universitygraduation(INE,2011).Because

basicliteracy(toknowhowtoreadandwrite)isarequisitetoobtain

aPortuguesedrivinglicence,illiteratepeoplearenotrepresented

inthestudysample.

2.2. Procedure

ThisstudywasapprovedbytheScientiﬁcBoardofFacultyof

Psychologyand EducationalSciences oftheUniversityof

Coim-braandtheMLTI.AlloftheethicalprinciplesoftheDeclarationof

Helsinkiwerefollowed.Priortoparticipation,thesubjectreceived

writteninformationregardingthepurposeofthestudy,and

writ-teninformedconsentwasobtained. Astheprotocolassessment

(3)

assessmentinPortugal(particularlyanon-roaddrivingtest),

sub-jectswereassuredthatresultswereconﬁdentialandhadnobearing

ontheirdrivinglicense.Because22(40%)oftheparticipantshadan

expireddrivinglicenseduetothewaitinglistandtime(2–3months

onaverage)betweenthephysician’sreferralandthedateof

assess-ment,anapplicationformwasalsosignedbytheparticipantand

anofﬁcialrepresentativeoftheMLTItopermittheparticipantsto

driveonapublicroad.Theevaluationswereperformed

individu-allyinasinglesession,andtheorderbetweenthecognitiveand

drivingtestingwascounterbalancedinthesample.Theresearch

protocolincludedamultimodalassessmentthatcollecteddataon

demographic,clinicalanddrivinghistory(throughthedriving

pro-cessandaninterview),visualfunction(farvisualacuity,chromatic

vision,andstereoscopicvision),cognitivefunctioning(ACE-R)and

drivingability.Althoughvisionwaspreviouslyscreenedby

physi-cians,visualfunctionwasexaminedpriortothewritteninformed

consenttoensuretheselectionofparticipantswithoutvisual

prob-lemsthatwouldaffectthecognitiveanddrivingoutcomes.

Theon-roaddrivingtestwascarriedoutbyanaccredited

pro-fessionalexaminerwhowasblindtotheparticipants’cognitiveand

personaldata.Theroadtestwasconductedona10-km

predeter-minedrouteincentralLisbon(Portugal)onthesamedayofthe

weekandduringanoff-peakperiod(between12hand14h).The

coursebeganinaresidentialareaasafamiliarisationperiodwith

theexperimentalvehicleanddriverexaminerandproceededto

localandarterialroads.Theroutewasoutlinedtoensurea

con-sistentlylowtomoderateleveloftrafﬁcconditionsandtoenable

theobservation of drivingbehaviour in differentdriving

situa-tions(e.g.,crossroads,roundabouts).Themeantotaltime ofthe

drivingtestwas50min.Thevehiclewasﬁttedwithmanual

trans-mission(themostfamiliarforallparticipants)andadual-brake

controlsystem.Theassessmentwasbasedonadrivingmeasure,

theDriving BehavioursObservation Grid (descriptionin

instru-mentssection), that wascompleted after theroad assessment.

Theexaminerprovidedaninitialstandardinstructionthatomitted

wordssuchas“assessment”or“examination”tominimisepotential

anxietythatmight produceanegative effectondriving

perfor-mance.Standardverbaldirectionalinstructionswerealsoprovided

duringthecourse,andtheexaminerintervenedphysically(using

thedual-brakeorsteeringwheel)orverballywhenrequiredfor

safety.Aninterraterreliabilitystudybetweenthedriving

exam-inerandaresearchpsychologist(inthefrontandrearpassenger

seat,respectively),whowaspreviouslytrainedondriving

obser-vation,wasconductedon24%ofthestudysample.Toensurethat

bothobserverswereblindtothecognitiveresults,theinterrater

assessmentsonlyincludedobservationsofparticipantswhohad

drivenbeforethecognitivetest.

2.3. Instruments

1.TheACE-Risabriefcognitivetestthatwasdevelopedtodetect

mildcognitiveimpairmentanddementia.FortheACE-Roriginal

reference,seeMioshietal.(2006).ItprovidesanMMSEscore

andﬁvesub-scoresthatrepresentcognitivedomains,namely

attentionandorientation(18pts),memory(26pts),letterand

categoryverbalﬂuency(14pts),language(26pts)and

visuospa-tialability(16pts).Thetotalscoreisobtainedbysummingallof

thesub-scores,rangingfrom0to100,ofwhichtheMMSEscore

(30pts)ispart.Higherscoresindicatebettercognitive

function-ing.ThePortugueseACE-Rversion(Firminoetal.,2008)used

inthepresentstudywasadaptedwithauthorisationandadvice

fromtheauthorsoftheoriginalversion.

2.TheDBOG(DrivingBehavioursObservation Grid)is adriving

performanceevaluation(unpublished)thatwasdevelopedfora

researchprojectontheneuropsychologicalassessmentofolder

drivers.Basedonaliteraturereviewofstudieswithroadtests

forpeoplewithcognitivedisabilities(Akinwuntanetal.,2003;

DeRaedtandPonjaert-Kristoffersen,2001;Fittenetal.,1995; Huntet al.,1997; Jankeand Eberhard,1998; Radford,2001),

thepresentversionwasdevelopedwiththeconsensusoftwo

independentexpertsondrivingexaminations.Todevelopavalid

drivingassessmentforolderpeople,thedrivinghabitsand

dif-ﬁculties thatpreviousresearchfoundfor thisgroup(Ferreira

etal.,2007)werealsotakenintoaccount.Thegridconsistsof

50itemsordrivingbehavioursdistributedacrossthefollowing

10categoriesorspeciﬁcscores:(a)pre-drivingchecks(3items),

asmirrors,seatbelt;(b)vehiclecontrol(7items),as

accelera-tor,footbrake;(c)visualsearch(3items),asfortrafﬁcsigns,for

potentialhazards;(d)communicationwithotherroadusers(2

items),asuseofindicator,responsetosignalsfromroadusers;

(e)drivingonurbanroads(15items),aslateralposition,

aware-nessofpedestrians;(f)drivingonthefreeway(6items),asspeed

choice,safetydistance;(g)turningatjunctions(4items),asgive

rightofway,trajectory;(h)drivingonroundabouts(4items),

aspositionforlanechoice,merging;(i)speciﬁcmanoeuvres(3

items), asturnin theroad,reverseparking; (j)otherspeciﬁc

behaviours (3items), asconﬁdence,self-regulationaccording

todeﬁcits.Eachdrivingitemisscoredonthefollowing3-point

scale:2=correctdrivingbehaviour,1=minorerror(not

affect-ingsafety)and0=majorerror(severeenoughtoaffectsafety).

Ageneralscoreisobtainedbysummingalloftheitems,ranging

from0to100,inwhichahigherscoreindicatesbetterdriving

performance.Basedonthefrequencyandseverityofobserved

behavioursduringthedrivingtask,theexaminerindicatesthe

ﬁnalqualitativeoutcomeoftheparticipant’sdrivingabilityas

safeor unsafe.The unsafeclassiﬁcation is appliedtodrivers

whocommitseveralsystematicerrorswithsufﬁcientseverityto

affectsafety(e.g.,lanechangeswithoutlooking,obstructsother

carsatcrossings,inappropriatebrake/accelerator).

2.4. Statisticalanalyses

Fordemographic, medical,drivingand cognitivedata,a

one-wayanalysisofvarianceorchi-squared(2₎_test_was_applied_to

comparethestudygroups.Discriminantanalyseswereperformed

toassesswhetherthepredictorvariables(independentvariables)

distinguishedbetweensafeandunsafedrivers(dependent

vari-able)onaroadtest.Receiveroperatingcharacteristic(ROC)curves

wereplottedtocomparetheareaunderthecurvebetweenthe

ACE-RandMMSEscores.ForthereliabilitystudyoftheDBOG,the

Pearson’scorrelationcoefﬁcientwasusedtoassesscorrelations

betweenspeciﬁcscoresandthegeneralscoreandtheinterrater

concordancebetweenallscores.Partialcorrelationswerealso

com-putedbetweenthedrivingmeasurescoresandACE-Routcomes.

AllanalyseswereconductedusingSPSSStatistics(version20.0,

Chicago,IL)withtheexceptionof theROCanalysis,which was

performedinMedCalc(version11.1,Mariakerke,Belgium).

3. Results

3.1. Comparisonofnon-cognitivevariablesbetweensafeand

unsafedrivers

AsshowninTable1,thesafeandunsafegroupsweresimilaron

thedemographic(age,genderandyearsofschooling),reasonfor

referral,visualfunction(farvisualacuityandstereoscopicvision)

and driving variables (years of driving experience, professional

driverexperienceanddrivingincidentsinprevioustwoyears).The

chromaticvision,notlistedinthetable,wasfoundasnormalinall

(4)

Table1

Demographic,medical,anddrivingcharacteristicsofthestudygroups.

Safe(n=25) Unsafe(n=25) pa Demographic Age(M,SD) 72.36,7.65 73.92,6.34 .436 Agegroups(65–74/75–88) 16/9 12/13 .254 Gender(male/female) 23/2 21/4 .384 Yearsofschooling(M,SD) 7.24,5.09 5.12,3.17 .084 Medical

Reasonforreferral(age/physical/neurological/psychiatric) 12/2/6/5 12/2/6/5 1.000 Farvisualacuity(M,SD) 0.67,0.15 0.60,0.14 .100 Stereoscopicvision(absent-reduced/average-superior) 9/16 8/17 .771 Drivingexperience

Yearsofdrivingexperience(M,SD) 44.56,11.85 39.60,12.83 .162 Professionaldriverexperience(yes/no) 6/19 4/21 .480 Drivingincidentsinprevioustwoyearsb_(yes/no) _7/18 _5/20 _.508

a_Differences_between_groups_were_tested_by_ANOVA,_with_the_exception_of_gender,_reason_for_referral,_professional_driver_experience_and_driving_incidents_in_previous

twoyears,whichwereevaluatedby2_with_one_degree_of_freedom.

b _Self-report_data_of_at-fault_automobile_crashes_and/or_state_records_of_moving_violations.

Table2

Mean,standarddeviationandrangevaluesofDBOGscoresbetweenstudygroups.

Safe(n=25) Unsafe(n=25) pvaluea

M SD Range M SD Range Generalscore 87.60 12.47 59–100 52.20 8.94 31–66 .000 Speciﬁcscores Pre-drivingchecks 5.60 0.87 3–6 4.28 1.51 1–6 .000 Vehiclecontrol 12.00 2.14 7–14 7.72 2.41 2–13 .000 Visualsearch 5.32 1.15 3–6 2.92 0.70 1–5 .000

Communicationwithotherroadusers 3.24 0.88 2–4 1.96 0.35 1–3 .000 Drivingonurbanroads 26.80 3.64 17–30 16.28 2.84 10–21 .000 Drivingonthefreeway 10.16 2.15 6–12 5.96 1.17 4–9 .000 Turningatjunctions 7.16 1.34 4–8 4.24 0.93 2–7 .000 Drivingonroundabouts 6.68 1.49 4–8 4.04 0.84 1–6 .000 Speciﬁcmanoeuvres 5.36 0.91 3–6 2.12 1.83 0–6 .000 Speciﬁcbehaviours 5.16 1.21 3–6 2.84 0.69 1–4 .000

a_Differences_between_groups_were_tested_by_ANOVA.

The results from the on-road assessment are presented in

Table 2. All scores from the DBOG were signiﬁcantly different

betweenthesafe and unsafegroups, demonstrating great

con-sistency between the objective observations registered by the

examinerandtheﬁnalqualitativeassessmentoftheparticipant’s

drivingability.Comparingtherangeofvaluesonthegeneralscore,

asmalloverlapwasobservedbetween59and66pointsinwhich

twodrivers(bothwith59points)wereeffectivelyclassiﬁedassafe.

3.2. ComparisonofACE-Routcomesbetweensafeandunsafe

drivers

Thereweresigniﬁcantdifferencesbetweengroups onACE-R

outcomes(Table3),withtheACE-Rtotalscorepresentingthe

great-estsigniﬁcance, despitetheoverlap of scores.The visuospatial,

ﬂuencyandlanguagesub-scoresweretheonlydiscriminative

sub-scores;however,theattention–orientationandmemorydomains

weremarginallyclosetosigniﬁcance.

3.3. DiscriminantanalyseswithACE-Rvariablestopredict

drivingability

Differentpredictorvariableswereanalysedindependentlywith

anentermethodtohighlighttheirrelativecontributioninthe

dis-criminantmodelincludingleave-one-outclassiﬁcation(Table4).

AllWilks’lambdasweresigniﬁcantatp<.05.

Theﬁrst twomodels (1and 2)analysed thecontributionof

anACE-RscoreorMMSEscorewithselectednon-cognitive

vari-ables(age,gender,yearsofschooling,farvisualacuityandyearsof

drivingexperience)inpredictingon-roaddrivingability.Theonly

Table3

Mean,standarddeviationandrangevaluesofACE-Routcomesbetweenstudygroups.

Safe(n=25) Unsafe(n=25) pvaluea

M SD Range M SD Range Scores ACE-R 78.28 10.95 47–94 66.48 14.10 43–93 .002 MMSE 26.48 2.80 18–30 24.48 3.69 18–30 .036 Sub-scores Attention–orientation 16.96 1.67 12–18 15.88 2.15 11–18 .053 Memory 16.96 5.78 3–26 14.20 4.66 5–24 .069 Fluency 8.28 2.87 3–14 5.76 3.05 1–12 .004 Language 22.44 3.00 15–26 19.32 5.14 11–26 .012 Visuospatial 13.64 2.34 9–16 11.32 2.89 7–16 .003 Note.ACE-R=Addenbrooke’sCognitiveExaminationRevised;MMSE=Mini-MentalStateExamination.

(5)

Table4

Percentageofsubjectscorrectlyclassiﬁedbytheclassiﬁcationfunctionwith leave-one-outcross-validation.

Predictormodel Sensitivity Speciﬁcity Accuracy 1.ACE-Rtotalscoreandnon-cognitive

variablesa

60 68 64

2.MMSEscoreandnon-cognitive variablesa

56 60 58

3.ACE-Rtotalscore 72 68 70 4.MMSEscore 56 76 66 5.Fivesub-scoresb ₆₀ ₆₄ ₆₂

6.Threediscriminatingsub-scoresc ₆₄ ₆₄ ₆₄

7.Eightdiscriminatingtasksfrom ACE-Rd

64 52 58

Note. Sensitivity=percentage of unsafe drivers correctly classified; Speci-ficity=percentageof safedrivers correctlyclassified; Accuracy=percentageof driverscorrectlyclassifiedbythemodel;ACE-R=Addenbrooke’sCognitive Exami-nationRevised;MMSE=Mini-MentalStateExamination.

a_ACE-R_total_score_or_MMSE_score,_years_of_schooling,_far_visual_acuity,_years_of

drivingexperience,genderandage(orderedbyabsolutesizeofcorrelationwithin functioninbothanalyses).

b_{Visuospatial,}_ﬂuency,_language,_{attention–orientation,}_and_memory_(ordered_by

absolutesizeofcorrelationwithinfunction).

c _{Visuospatial,}_ﬂuency_and_language_(ordered_by_absolute_size_of_correlation

withinfunction).

d _Phonemic_verbal_ﬂuency,_retrograde_memory,_pentagons,_naming₁₀_pictures,

picturescomprehension,clock,delayedmemory,andcube(orderedbyabsolutesize ofcorrelationwithinfunction).

discriminativevariableinthesemodelswasthecognitivescore,

andthevariableagehadthelowestcontributiontothe

classiﬁca-tionfunctions.Thenexttwoanalyses(models3and4)withoutthe

non-cognitivevariablesdemonstratedsuperiorclassiﬁcation

accu-racy,withalowerpercentageofclassiﬁcationerrors.Theaccuracy

ofpredictivemodelsofdrivingabilitywithanACE-RscoreorMMSE

scoredecreasedwhennon-cognitivevariableswereaccountedfor

inthediscriminantfunctionanalyses.Ratherthancontributingto

thediscriminationbetweengroups,thesevariablesseemedto

pro-ducerandomerrorsornoise.

TheACE-Rscoremodeldemonstratedhigherclassiﬁcation

accu-racy (70%; 35 out of 50 drivers correctly classiﬁed), with 72%

sensitivity(18/25unsafecorrectlyclassiﬁed),68%speciﬁcity(17/25

safecorrectlyclassiﬁed),69.2%positivepredictivevalueand70.8%

negativepredictivevalue.TheMMSEscoremodelexhibitedminor

sensitivity(56%;14/25unsafecorrectlyclassiﬁed),highspeciﬁcity

(76%;19/25safecorrectlyclassiﬁed),andaﬁnalaccuracyof66%

(33/50correctlyclassiﬁed),producingpositiveandnegative

pre-dictivevaluesof70%and63.3%,respectively.

Concerning the ACE-R cognitive domains (models 5 and 6),

thevisuospatial, ﬂuencyand languagesub-scoresweretheonly

sub-scoresthatsigniﬁcantlydiscriminatedthetwo groups,with

ahigherclassiﬁcationaccuracy(64%) thanthepredictivemodel

with ﬁve sub-scores (62%), including the attention–orientation

andmemorydomains. Theﬁnalmodel (model7)thatincluded

eightdiscriminatingtasksbetweenthestudygroups(retrograde

memory,phonemicverbalﬂuency,naming10 pictures,pictures

comprehension,pentagons,cube,clockanddelayedmemory)only

correctlyclassiﬁed58%ofthedrivers.

Tounderstandthecontributionofeach ACE-Rdiscriminating

taskinaccuratelypredictingsafeandunsafedrivers,discriminant

analysesweremadeindependently(Table5).Thephonemicverbal

ﬂuencytaskrepresentedthemostefﬁcientmodel,withanaccuracy

of74%(37/50driverscorrectlyclassiﬁed).Thelanguagetasks

(nam-ingtenpicturesandpicturescomprehension)andavisuospatial

task(pentagons)werethemostefﬁcientindetectingsafedrivers

(speciﬁcityrangingfrom80to92)butwerelessrobustindetecting

unsafedrivers(sensitivityrangingfrom48to56).Other

visuospa-tialtasks(copyingacubeandclockdrawing)andmemorytasks

Table5

Sensitivity,speciﬁcityandaccuracyoncross-validationresultsofACE-R discrimi-natingtasksbetweensafeandunsafedrivers.

Predictormodel Sensitivity Speciﬁcity Accuracy

Retrogradememory 56 76 66

Phonemicverbalﬂuency 80 68 74

Namingtenpictures 48 80 64

Picturescomprehension 48 92 70

Pentagons 56 80 68

Cube 64 64 64

Clock 64 64 64

Delayedmemory 68 64 66

(retrogradeand delayedmemory)alsoefﬁcientlypredictedsafe

andunsafedrivers.

3.4. ComparisonofROCcurvesbetweentheACE-Rscoreand

MMSEscore

The area under the ROC curve of the ACE-R score was .75.

This score is signiﬁcantly higher than an AUC of .05 for no

discrimination(z=3.46;p<.001) andrepresentsacceptable

dis-criminativepower(0.7≤ROC<0.8).TheareaundertheROCcurve

of theMMSE score was.66, which is alsohigher than an AUC

of.05fornodiscrimination(z=2.05;p=.040),butwithlow

dis-crimination (0.5<ROC<0.7). The AUC of the ACE-R score was

considerably largerthan theAUCof theMMSE score,although

these differences did not reach signiﬁcance (z=1.68; p=.093)

(Fig.1).

3.5. ReliabilityofDBOG

ThehomogeneityoftheDBOGwascalculatedwithcorrelations

betweentheDBOGspeciﬁcscoresandtheDBOG generalscore,

whichrangedbetween0.516(p<.086)and0.963(p<.000).With

theexceptionof“pre-drivingchecks”,allofthespeciﬁcscoreshad

astrongcorrelation(r>0.7)withthegeneralscore.

InregardstotheresultsoftheDBOG’sinterraterreliabilitystudy,

thedichotomousqualitativeoutcome(safe/unsafe)ofthe

partici-pants’drivingabilitywas100%identicalbetweenobservers.The

correlationsbetweenquantitativescoreswere:pre-drivingchecks

(r=0.06,p=.847),vehiclecontrol(r=0.79,p=.002),visualsearch

(r=0.89,p=.000),communicationwithotherroadusers(r=0.87,

p=.000),drivingonurbanroads(r=0.94,p=.000),drivingonthe

freeway(r=0.83,p=.001),turningatjunctions(r=0.93,p=.000),

driving on roundabouts (r=0.90, p=.000), speciﬁc manoeuvres

(r=0.98,p=.000),speciﬁcbehaviours(r=0.91,p=.000),and

gen-eralscore(r=0.97,p=.000).

(6)

Table6

Partialcorrelationsbetweenon-roaddrivingandACE-Routcomes(aftercontrollingforvisualacuity).

Scores Sub-scores

ACE-R MMSE Attention/orientation Memory Fluency Language Visuospacial Speciﬁcscores

Pre-drivingchecks

Vehiclecontrols .44** _.33* _.39** _.32* _.40** _.33* _.33*

Visualsearch .51** _.36* _.34* _.40** _.46** _.40** _.37*

Communicationwithotherroadusers .39** _.39** _.32* _.30*

Drivingonurbanroads .47** _.31* _.32* _.37* _.44** _.34* _.35*

Drivingonthefreeway .40** _.34* _.35* _.35*

Turningatjunctions .35* _.35* _.35* Drivingonroundabouts .38** _.37** _.30* _.36* Speciﬁcmanoeuvres .38** _.37** _.29* Speciﬁcbehaviours .34* _.37* _.33* Generalscore .45** _.30* _.31* _.34* _.42** _.33* _.36* * _p_<_.05. ** _p_<_.01.

3.6. AssociationsbetweenDBOGandACE-Routcomes

Todeterminewhich roadtestvariablesfromtheDBOGwere

related to theACE-R outcomes, partialcorrelations were

com-puted with visual acuity as a covariate, as this variable was

correlated witha number of driving scores(rranged from .15

to .35, p<.05). The “pre-driving checks” was the only speciﬁc

scorewithoutanassociationwithcognitiveoutcomes.The

ACE-Rscoreandﬂuencyandvisuospatialsub-scoreswereassociated

withallotherdrivingscores.Comparatively,theMMSEscoreand

theattention–orientationsub-scorewereonlyrelatedwiththree

speciﬁcscores,vehiclecontrol,visualsearch,anddrivingonurban

roads,whichwereassociatedwithallcognitivedomains.The

mem-ory sub-scorewas related to “communication with other road

users”and “drivingonthefreeway”, and languagedomainwas

associatedwith“drivingonroundabouts”.Allofthecorrelations

weremoderate(Table6).

4. Discussion

Thepresent studyaimedtoanalysethe potentialvalueof a

briefcognitivetest,theACE-R,inpredictingdrivingabilityinolder

drivers.Thestudygroups,driversclassiﬁedassafeorunsafeon

a real driving test, had similar demographic and speciﬁc

clini-calvariables.Thisallowedanapproachtotheneuropsychological

functionsassessed bytheACE-R withthegreatestinﬂuenceon

drivingperformance.Thediscriminativevariablesbetweengroups

werethecognitiveanddrivingresults,whichmayrepresentthe

ﬁrstsignthatthisbriefcognitivetestisusefulinpredictingon-road

drivingability.

Amajorﬁndingwasthehigherclassiﬁcationaccuracyofthe

ACE-RtotalscoreinrelationtotheMMSEscore(includedinthe

ACE-R),particularlyindetectingunsafedrivers.Theoverall

accu-racyoftheACE-Rwas70%,withbalancedsensitivity(72%) and

speciﬁcity(68%).Comparatively,theoverallaccuracyoftheMMSE

scorewasnotonlylower(66%),butalsopresentedamajor

trade-offbetween sensibility(56%) andspeciﬁcity (76%), producing a

greaterpercentageoffalsenegatives(unsafeincorrectlyidentiﬁed

assafedrivers).Thus,theMMSEscorewaslessusefulfor

detect-ingunsafedrivers.TheseresultssuggestthatMMSEscoresmay

beinappropriateforthescreeningofdrivingimpairmentsina

pri-marycaresampleofolderpeople.TheresultsfromtheROCanalyses

demonstratedthefollowingtrend:theACE-Rhadgreater

discrimi-nantabilityinthepredictionofsafe/unsafedriversthantheMMSE,

althoughthedifferencesdidnotreachstatisticalsigniﬁcance.The

lackofstatisticalsigniﬁcancemaybeduetothelowerpowerofthe

AUCcomparisontestgiventhelimitedsamplesize.TheACE-Rscore

seemstopredictdrivingabilitybetterthantheMMSEscore.

How-ever,ofnote,theACE-Rscoredoesnotpresentsufﬁcientlyhigh

levelsof sensibilityand speciﬁcity(conventionallyboth greater

than80%)tobeusedasascreeningtesttopredictunsafedrivers.

The visuospatial, ﬂuency and language sub-scores were the

only sub-scores that were predictive of driving ability, while

attention–orientationandmemorydomainsweremarginallyclose

tostatisticalsigniﬁcance.Considerable researchhasshown that

visuospatialability measuresare strongly correlated with

driv-ingability(Regeretal.,2004).Concerningthevisuospatialtasks

includedontheACE-R,apoorpentagondesign(copyingof

over-lappingpentagons,fromtheMMSE)waspredictiveofself-reported

drivingdifﬁculties(Galloetal.,1999)andautomobilecrashesand

movingviolations(Marottolietal.,1994),copyingacubewasa

strongpredictorofcrashes(Johanssonetal.,1996),andclock

draw-ingwaspredictiveof real(De RaedtandPonjaert-Kristoffersen,

2001)andsimulateddrivingtasks(Freundetal.,2008)inactive

olderdrivers.

Previousresearchondrivingabilitieswaslessconclusivewith

respectto theutility of verbalﬂuency tasks, mainly

conceptu-alised as measures of executive functionsand language(Lezak

et al., 2004; Strauss et al., 2006). For instance, a category

ﬂu-ency task was predictive of simulated driving performance in

patientswithAlzheimer’sdisease(AD)(Rebok etal.,1994), and

theControlled Oral Word Association Test (COWA)was a

pre-dictorofon-roaddrivingperformanceinpatientswithearlyAD

(Ucetal.,2004).However,theWordFluencyTest(Ducheketal.,

1998; Huntet al.,1993)for phonemic associationand

Genera-tiveNaming(Whelihanetal.,2005)forsemanticassociationwere

notrelatedtoon-roaddrivingperformanceinpatientswithearly

cognitivedeclineorAD.Ofnote,theverbalﬂuencytaskswere

pri-marilystudiedindriverswithAD,and thepresent studyfound

thattheﬂuencysub-scoreoftheACE-Rwasusefulinpredicting

drivingability among a clinicalsample of olderadults referred

fromprimarycaresettings,withanunspeciﬁc(“age”)ordifferent

diagnosis.

ThelanguagedomainoftheACE-Rwasalsopredictiveof

driv-ingability.Althoughlanguagefunctionshavebeenshowntobe

irrelevantfordrivingbehaviour(e.g.,Golperetal.,1980)andhave

notbeenrecommendedaspartofaneuropsychologicalbatteryfor

driving(McKenna,1998),itisreasonabletoexpectthatlanguage

deﬁcitsmaybeindicativeofaglobalcognitivedeclinethat

compro-misesdrivingperformance.Forinstance,theBostonNamingTest

hasshownpositiveassociationswithon-roaddrivingperformance

indementiacases(Ducheketal.,1998;Huntetal.,1993).

Simi-larly,thecurrentstudyfoundthatthelanguagesub-scoreandother

cognitivesub-scoresoftheACE-Rwereassociatedwiththesame

(7)

Farfromexpected,theattention–orientationdomain,included

in the MMSEscore, did not sufﬁciently discriminate thestudy

groups.Theroleofattentionalskillsindrivingiswelldocumented

(Balletal.,1993;ParasuramanandNestor,1991).However,this

cognitivedomainoftheACE-Rdoesnotspeciﬁcallyassess

atten-tion,asmorethanhalfofthisscoreisrelatedtoorientationitems.

TobetterunderstandwhethertheACE-Rissufﬁcientlyrobustin

theassessmentofattention,afurtherstudymustdistinguish

sub-scoresfororientationandattentionitems.

Thememorydomainshowedtheleastsigniﬁcantdifferences

betweensafeandunsafedrivers.Studieshaveshownarelationship

betweenmemorydeﬁcitsandtheincreasedriskofdrivingerrors,

especiallyinpatientswithaseverecognitiveimpairment(Hunt

etal.,1993).Memorytestsarequitesensitivetooverallcognitive

impairment,asthepresenceofmemorydeﬁcitsisakeycriterion

inthediagnosisof mildcognitiveimpairment ordementia. For

instance,morethanone-quarteroftheACE-Rscorepertainstothe

memorydomain.Thepresentresultsdemonstratedassociations

betweenthememorydomainanddrivingperformance;however,

theseassociationswerelessstrikingthanthosefortheﬂuencyand

visuospatialdomains(associatedwithalldrivingscores).Thus,the

ACE-Rmemoryscorewasnotthemostfacevalidcognitivedomain

foron-roaddriving.Furthermore,thepresenceofmemory

impair-mentsmayinteractwithimpairmentsinothercognitivedomains

relevanttodriving.Toisolatetheeffectsofmemoryimpairmentin

driving,Andersonetal.(2007)foundthatdriverswithsevere

amne-siaduetobilateralhippocampallesionswerenotimpairedinmost

aspectsofdrivingperformance(knowledgeofdrivingrules,safety

procedures,androadsignmeaningwerenormal),butpresented

somedifﬁcultiesinfollowingroutedirectionsontheroad.Astudy

fromGraceetal.(2005)that comparedneuropsychological and

drivingperformanceinADandParkinsonDiseasepatientsfound

thatseverememoryimpairmentintheADgroup(i.e.,a cortical

dementiawithimpairedlearning,retrievalandrecognition

mem-ory)didnotrelatetopoordriving,butmemoryimpairmentinthe

unsafeParkinsondiseasegroup(i.e.,asubcorticaldementiawith

impairedretrievalbutrelativelypreservedrecognition)reﬂected

theinﬂuenceofexecutivedysfunctiononmemory.Thesedata

cor-roboratethepresentﬁndingsthataneuropsychologicaltestwitha

largeexecutivecomponentmaybemorerobustfordriving-related

functionsthanmemorytasks.

Anabbreviatedform oftheACE-R witheight discriminating

tasksbetweensafeandunsafedriverswasnotsuccessful(58%of

accuracy)inpredictingdrivingability.Thepredictormodelwith

theACE-Rscorewasmoreefﬁcient,possiblybecauseitcoversmore

driving-relatedcognitivedataorinformationandthereforeprides

areasonablecompromisebetweenthedynamicsofcognitive

func-tioningand drivingperformance.From thediscriminatingtasks

betweenthesafeandunsafegroups,onlythepentagontaskwas

partoftheMMSEscore.Incontrast,severaldiscriminatingtasks

informedtheACE-Rscore,whichcontributednotonlytoagreater

accuracy,butalsotoamorebalancedmodelinthecorrect

classiﬁ-cationofdrivers.

Weaimedtounderstandtheabilityofeachdiscriminativetask

fromtheACE-Rtocorrectlyclassifydrivers.Retrogradememory,

namingtenpictures,picturescomprehensionandpentagonswere

moretailoredfordetectingsafedrivers(allwithaspeciﬁcity>75

butwithlowsensitivity).Phonemicverbalﬂuencywasdistinctive

fordetectingunsafedrivers,withhighsensitivityandoverall

accu-racy.Finally,cube,clockanddelayedmemorywerebalancedtasks

forclassifyingsafeandunsafedrivers(whichcouldbefairfora

pre-dictormodel),butalonewerenotsufﬁcientlyaccurate(accuracy

rangedbetween64and66).

OftheeightdiscriminatingitemsoftheACE-R,phonemic

ver-balﬂuencywasthebestclassiﬁerof drivingability,particularly

for detecting unsafe older drivers. It was also associated with

all driving-speciﬁcscores in thepresent study. Thisis a

sensi-tive task for frontal lobe functions (Alvarez and Emory, 2006)

andhencetofrontotemporaldementia,whereascategoryﬂuency

isa markerof semanticmemorybreakdowntypicalin AD(e.g.,

Mathuranathetal.,2000).Phonemicverbalﬂuencyisconsidered

valuableindetectingcognitivealterationsintheagedgivenits

sta-bilitythroughouttheageingprocess(Brickmanetal.,2005),which

mayofferdiagnosticutilityintheassessmentofolderdrivers.

How-ever,itisimportanttonotethatoneitemfromtheACE-Rdoesnot

tellthewhole“story”aboutdrivingability,asvisuospatial

abili-tiesandvisualattention(modalitiesnotexaminedintheACE-R)

arealsocrucialforsafedriving(Regeretal.,2004;Mathiasand

Lucas,2009).Duetothecognitivecomplexityanddynamics

intrin-sictodriving,asingletaskisinsufﬁcientfordeterminingwhether

anindividualshoulddrive(Rizzo,2011).

ConcerningtheassociationsbetweentheDBOGandACE-R

out-comes,theresultssuggestedthatglobalcognitivescreeningand

speciﬁccognitivedomainsof executivefunctionsand

visuospa-tialabilitieswereassociatedwithalldrivingskills,stressingthat

deﬁcitsinthesefactorsrepresentimpaireddrivingability.

Compar-atively,theremainingdomainsandtheMMSEscoredidnotappear

(again)asthemostfacevalidACE-Routcomesforon-roaddriving

performance.

ToimprovetheaccuracyoftheACE-Rinpredictingdriving

abil-ity, a speciﬁcmodel ortest ratio witha cut-off point couldbe

studiedinthefuturetodifferentiatesafeandunsafedrivers.Sucha

studywouldbesimilartotheworkdevelopedbyMathuranathetal.

(2000)todifferentiateADandfrontotemporaldementiaintheﬁrst

ACEpaper.However,theACE-Rmustbestudiedwithother

comple-mentaryteststhathavesystematicallypredicteddrivingoutcomes,

suchasmeasuresofvisualattention(e.g.,DotCounting,Trail

Mak-ingTest–AandUFOV),executivefunctioning(e.g.,TrailMaking

Test–B,KeySearchandRuleShift)andvisuospatialabilities(e.g.,

SquareMatricesDirections,RoadSignRecognition,WAIS-IIIBlock

Design)(cf.,Balletal.,1993;Ferreiraetal.,2010;Lincolnetal.,

2010;McKennaandBell,2007).

Thestudyofpredictingdrivingabilityusingneuropsychological

testsimpliesareliableandvaliddrivingoutcome.Theconsistent

resultsbetweentheobservationsregisteredbytheexaminerand

theﬁnalqualitativeoutcomeoftheparticipants’drivingabilityis

suggestiveoftheobjectivityoftheDBOG(whenusedbyadriving

specialist)toassessdrivingperformanceinolderdrivers.However,

a slightoverlap inthegeneralscore ofthe studygroups raises

thequestion ofwhetherscoresshould besummedtoan

over-allscoreorclassiﬁcationsshouldbebasedonaﬁnalqualitative

outcome.Bothcriteriawereconsideredinthepresentstudy.The

overallimpressionoftheexaminerwasbasedontheDBOGasan

operationalisedmeasureofdrivingbehaviour,withoutneglecting

theexaminer’sexperienceand expertisetosubjectivelyclassify

adriverassafeorunsafe.Ofnote,anoverallscoremaybemore

affectedbyminorerrors(notaffectingsafety)or“badhabits”(cf.,

Selanderetal.,2011),whichshouldbedistinguishedandnot

con-sideredbyatrainedobserverforanunsafedrivingclassiﬁcation.

TheresultsfromthereliabilityoftheDBOGalsoshowedthatthis

drivingmeasure wasobjectiveand stablewhen usedby

differ-enttrained observers.However,onespeciﬁc score,“pre-driving

checks”(adjustmentofseat,mirrorsandputtingontheseatbelt),

wasnotsufﬁcientlyconsistentduringobservations. Wesuggest

that these preliminary driving behaviours may deserve careful

attentioninafurtherinvestigationwiththisinstrument.

Theresultsofthepresentstudyprovidesupportforthepotential

utilityoftheACE-Rasascreeninginstrumentforolderdrivers.We

arguethatitmightbesimpleandlesstimeconsumingforprimary

carephysicianstoadministratea briefscreening toolfor

cogni-tiveimpairmentintheﬁrststageoftheﬁtnesstodriveassessment

(8)

consideringthediversityofneuropathologyassociatedwithaging,

whichisnotconsistentlydiagnosedorwell-documentedinprimary

healthcaresettingsandreferrals(cf.,McKennaandBell,2007).A

screeningtoolmustaidcliniciansindeterminingwhethera

per-sonshouldbereferredforamoredetailedevaluation(Wolfeand

Clark,2012).Thus,thepredictionofunsafedriversduetocognitive

impairmentshouldbeasecondstageoftheassessmentprocess

thatisconductedinspecialisedcentres.Suchanevaluationmust

focusonin-depthanalysesofneuropsychologicalfunctionswith

validatedteststhatrelateto(desirably)variousdrivingmeasures,

suchasdrivingperformanceandaccidentinvolvement.Inthe

spe-cialisedcentresofmanycountries,aroadtestisconsideredasan

additionalandcomplementarysourceofevidenceofdrivingsafety.

Somelimitationsofthecurrentstudyshouldbenoted.First,the

samplesizeissmall;therefore,thepresentﬁndingsrequire

fur-therindependentvalidationwithalargersample.In Portugal,a

roaddrivingassessmentisnotprovidedinaﬁtnesstodrive

assess-ment,whichleadstoconstraintsindeveloping afullevaluation

thatincludeson-roaddriving.Notwithstandingtheuseofthe“gold

standard”andvalidityoftheexternalcriterion,itispossiblethat

somebiasmayhaveoccurredintheclassiﬁcationofdrivers.The

presenceofundiagnosedphysical disordersiscommoninolder

people,andsomeexplicitmotordysfunctionsmayslightlyharm

drivingperformance.Despitetheselimitations,nopreviousstudies

analysedtherelationshipbetweentheACE-Randreal-world

driv-ingperformance.Thecurrentﬁndingssupportthepotentialvalue

oftheACE-Rin predictingdrivingability,asitismoreeffective

thananMMSEscore.Cognitivedomainsofexecutivefunctionsand

visuospatialabilitieswereparticularlyrelatedtodriving

perfor-mance.Thisemergingcognitivetestmaywarrantadditionalstudy

inrelationtodrivingmeasures,butitseemsasthoughtheACE-R

couldbeincludedinrevisedpsychologicalassessmentprotocolsfor

determiningdrivingﬁtnessinolderadults.

Conﬂictsofinterest

Allauthorsofthisstudydeclarethattherearenoconﬂictsof

interest.

Acknowledgements

WewouldliketothankthepartnershipbetweentheFaculty

of Psychology and Educational Sciences from the University of

CoimbraandtheMobilityandLandTransportsInstitutefor

provid-ingconditionstomakethisstudypossible.WealsothankSérgio

MarquesfromtheAutomobileClubofPortugalforthetimeand

expertiseincarryingoutthedrivingassessments.

Theauthorswouldalsoliketothanktheanonymousreferees

fortheircommentsandsuggestions.

Thiswork was supported by the Fundac¸ão para a Ciência e

Tecnologia [PortugueseFoundation for Scienceand Technology]

throughofaPhD fellowship(SFRH/BD/27255/2006)awardedto

theﬁrstauthor.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in

theonlineversion,atdoi:10.1016/j.aap.2012.03.036.

References

Akinwuntan,A.E.,Deweerdt,W.,Feys,H.,Baten,G.,Arno,P.,Kiekens,C.,2003. Relia-bilityofaroadtestafterstroke.ArchivesofPhysicalMedicineandRehabilitation 84(12),1792–1796.

Alexopoulos,P.,Ebert,A.,Richter-Schmidinger,T.,Schöll,E.,Natale,B.,Aguilar,C.A., Gourzis,P.,Weih,M.,Perneczky,R.,Diehl-Schmid,J.,Kneib,T.,Förstl,H.,Kurz,A., Danek,A.,Kornhuber,J.,2010.ValidationoftheGermanrevisedAddenbrooke’s

CognitiveExaminationfordetectingmildcognitiveimpairment,milddementia inAlzheimer’sdiseaseandfrontotemporallobardegeneration.Dementiaand GeriatricCognitiveDisorders29(5),448–456.

Alvarez,J.,Emory,E.,2006.Executivefunctionandthefrontallobes:ameta-analytic review.NeuropsychologyReview16(1),17–42.

Amick,M.M.,Grace,J.,Ott,B.R.,2007.Visualandcognitivepredictorsofdriving safetyinParkinson’sdiseasepatients.ArchivesofClinicalNeuropsychology22 (8),957–967.

Anderson,S.W.,Rizzo,M.,Skaar,N.,Stierman,L.,Cavaco,S.,Dawson,J.,Damásio,H., 2007.Amnesiaanddriving.JournalofClinicalandExperimental Neuropsychol-ogy29(1),1–12.

Bak,T.H.,Mioshi,E.,2007.AcognitivebedsideassessmentbeyondtheMMSE:the Addenbrooke’sCognitiveExamination.PracticeNeurology7(4),245–249. Ball,K.,Owsley,C.,Sloane,M.E.,Roenker,D.L.,Bruni,J.R.,1993.Visualattention

problemsasapredictorofvehiclecrashesinolderdrivers.Investigative Oph-thalmologyandVisualScience34(11),3110–3123.

Brickman,A.M.,Paul,R.H.,Cohen,R.A.,Williams,L.M.,Macgregor,K.L.,Jefferson, A.L.,Tate,D.F.,Gunstad,J.,Gordon,E.,2005.Categoryandletterverbalﬂuency acrosstheadultlifespan:relationshiptoEEGthetapower.ArchivesofClinical Neuropsychology20(5),561–573.

Carr,D.B., Ott,B.R., 2010. Theolderadult driverwithcognitive impairment: “It’saveryfrustratinglife”.TheJournaloftheAmericanMedical303(16), 1632–1641.

Daigneault,G.,Joly,P.,Frigon,J.-Y.,2002.Executivefunctionsintheevaluationof accidentriskofolderdrivers.JournalofClinicalandExperimental Neuropsy-chology24(2),221–238.

DeRaedt,R.,Ponjaert-Kristoffersen,I.,2001.Shortcognitive/neuropsychological testbatteryforﬁrst-tierﬁtness-to-driveassessmentofolderadults.Clinical Neuropsychology15(3),329–336.

Duchek,J.M.,Hunt,L.,Ball,K.,Buckles,V.,Morris,J.C.,1998.Attentionanddriving performanceinAlzheimer’sdisease.JournalsofGerontology.SeriesB, Psycho-logicalSciencesandSocialSciences53B(2),P130–P141.

Dudas,R.B.,Berrios,G.E.,Hodges,J.R.,2005.TheAddenbrooke’sCognitive Exam-ination(ACE)inthedifferentialdiagnosisofearlydementiasversusaffective disorder.AmericanJournalofGeriatricPsychiatry13(3),218–226.

EuropeanRoadSafetyObservatory,2006.OlderDrivers.,http://www.erso.eu. Ferreira,I.S.,Marmeleira,J.F.,Godinho,M.B.,Simões,M.R.,2007.Cognitivefactors

andgenderrelatedtoself-reporteddifﬁcultiesinolderdrivers.In: Proceed-ingsofthe4thInternationalDrivingSymposiumonHumanFactorsinDriver Assessment,TrainingandVehicleDesign,pp.511–518.

Ferreira,I.S.,Simões,M.R.,Marques,S.G.,Figueiredo,M.N.,Marmeleira,J.F.,2010. Neuropsychologicalassessment ofolderdrivers:review andsynthesis. In: SelectedProceedingsofthe12thWorldConferenceonTransportResearch(ID 02531),pp.1–24.

Firmino,H.,Simões,M.,Pinho,S.,Cerejeira,J.,Martins,C.,2008.Avaliac¸ão Cogni-tivadeAddenbrooke.ExperimentalPortugueseVersionoftheAddenbrooke’s CognitiveExamination–Revised(ACE-R;J.Hodges,E.Mioshi,2005®_).

Man-ual.PsychologicalAssessmentService.FacultyofPsychologyandEducational SciencesofUniversityofCoimbra.

Fitten,L.J.,Perryman,K.M.,Wilkinson,C.J.,Little,R.J.,Burns,M.M.,Pachana,N., Mervis,J.R.,Malmgren,R.,Siembieda,D.W.,Ganzell,S.,1995.Alzheimerand vasculardementiasanddriving.TheJournaloftheAmericanMedical273(17), 1360–1365.

Folstein,M.F.,Folstein,S.E.,Mchugh,P.R.,1975.“Mini-mentalstate”.Apractical methodforgradingthecognitivestateofpatientsfortheclinician.Journalof PsychiatricResearch12(3),189–198.

Freund,B.,Colgrove,L.A.,Petrakos,D.,Mcleod,R.,2008.Inmycarthebrakeison theright:pedalerrorsamongolderdrivers.AccidentAnalysisandPrevention 40(1),403–409.

Gallo,J.J.,Rebok,G.W.,Lesikar,S.E.,1999.Thedrivinghabitsofadultsaged60years andolder.JournaloftheAmericanGeriatricsSociety47(3),335–341. Golper,L.A.,Rau,M.T.,Marshall,R.C.,1980.Aphasicadultsandtheirdecisionson

driving:anevaluation.ArchivesofPhysicalMedicineandRehabilitation61(1), 34–40.

Grace,J.,Amick,M.M.,D’Abreu,A.,Festa,E.K.,Heindel,W.C.,Ott,B.R.,2005. Neu-ropsychologicaldeﬁcitsassociatedwithdrivingperformanceinParkinson’sand Alzheimer’sdisease.JournaloftheInternationalNeuropsychologicalSociety11, 766–775.

Hamilton,F.,Rochester,L.,Paul,L.,Rafferty,D.,O’Leary,C.,Evans,J.,2009. Walk-ingandtalking:aninvestigationofcognitive-motordualtaskinginmultiple sclerosis.MultipleSclerosis15(10),1215–1227.

Hunt,L.,Morris,J.C.,Edwards,D.,Wilson,B.S.,1993.Drivingperformancein per-sonswithmildseniledementiaoftheAlzheimertype.JournaloftheAmerican GeriatricsSociety41(7),747–752.

Hunt,L.A.,Murphy,C.F.,Carr,D.,Duchek,J.M.,Buckles,V.,Morris,J.C.,1997. Reliabil-ityoftheWashingtonUniversityRoadTest:aperformance-basedassessment fordriverswithdementiaoftheAlzheimertype.ArchivesofNeurology54(6), 707–712.

InstitutoNacionaldeEstatística,2011.PortaldeEstatísticasOﬁciais(Portuguese OfﬁcialStatistics).,http://www.ine.pt.

Iverson,D.J.,Gronseth,G.S.,Reger, M.A.,Classen,S.,Dubinsky,R.M.,Rizzo,M., 2010.Practiceparameterupdate:evaluationandmanagementofdrivingrisk indementia.Neurology74(16),1316–1324.

Janke, M.K.,Eberhard, J.W., 1998. Assessing medicallyimpaired olderdrivers in a licensing agency setting. Accident Analysis and Prevention 30 (3), 347–361.

(9)

Johansson,K.,Bronge,L.,Lundberg,C.,Persson,A.,Seideman,M.,Viitanen,M.,1996. Canaphysicianrecognizeanolderdriverwithincreasedcrashriskpotential? JournaloftheAmericanGeriatricsSociety44(10),1198–1204.

Kwak,Y.T.,Yang,Y.,Kim,G.W.,2010.KoreanAddenbrooke’sCognitiveExamination revised(K-ACER)fordifferentialdiagnosisofAlzheimer’sdiseaseand subcorti-calischemicvasculardementia.Geriatrics&GerontologyInternational10(4), 295–301.

Lezak,M.,Howieson,D.,Loring,D.,2004.NeuropsychologicalAssessment,4thed. OxfordUniversityPress,NewYork.

Lincoln,N.B.,Taylor,J.L.,Vella,K.,Bouman,W.P.,Radford,K.A.,2010.A prospec-tivestudyofcognitiveteststopredictperformanceonastandardisedroadtest inpeoplewithdementia.InternationalJournalofGeriatricPsychiatry25(5), 489–496.

Marcotte,T.,Scott,J.,2009.Neuropsychologicalperformanceandtheassessmentof drivingbehaviour.In:Grant,I.,Adams,K.(Eds.),NeuropsychologicalAssessment ofNeuropsychiatricandNeuromedicalDisorders.,3rded.OxfordUniversity Press,NewYork,pp.652–687.

Marottoli,R.A.,Cooney,L.M.,Wagner,D.R.,Doucette,J.,Tinetti,M.E.,1994.Predictors ofautomobilecrashesandmovingviolationsamongelderlydrivers.Annalsof InternalMedicine121(11),842–846.

Mathew,R.,Bak,T.H.,Hodges,J.R.,2011.Screeningforcognitivedysfunctionin cor-ticobasalsyndrome:utilityofAddenbrooke’scognitiveexamination.Dementia andGeriatricCognitiveDisorders31(4),254–258.

Mathias,J.L.,Lucas,L.K.,2009.Cognitivepredictorsofunsafedrivinginolderdrivers: ameta-analysis.InternationalPsychogeriatrics21(4),637–653.

Mathuranath,P.S.,Nestor,P.J.,Berrios,G.E.,Rakowicz,W.,Hodges,J.R.,2000.Abrief cognitivetestbatterytodifferentiateAlzheimer’sdiseaseandfrontotemporal dementia.Neurology55(11),1613–1620.

McKenna,P.,1998.Fitnesstodrive:aneuropsychologicalperspective.TheJournal ofMentalHealth7(1),9–18.

McKenna,P.,Bell,V.,2007.Fitnesstodrivefollowingcerebralpathology:the Rook-woodDrivingBatteryasatollforpredictingon-roaddrivingperformance. JournalofNeuropsychology1(1),85–100.

Mioshi,E.,Dawson, K.,Mitchell,J.,Arnold,R.,Hodges, J.R.,2006.The Adden-brooke’sCognitiveExaminationRevised(ACE-R):abriefcognitivetestbattery for dementia screening. International Journal of Geriatric Psychiatry 21, 1078–1085.

Morgan,E.E.,Heaton,R.K.,2009.Neuropsychologyinrelationtoeveryday func-tioning. In:Grant, I., Adams,K. (Eds.),Neuropsychological Assessmentof NeuropsychiatricandNeuromedicalDisorders.,3rded.OxfordUniversityPress, NewYork,pp.632–651.

OrganisationforEconomicCo-operationandDevelopment,2001.Ageingand Trans-port:MobilityandSafetyIssues.OECDPublications,Paris.

Ott,B.R.,Anthony,D.,Papandonatos,G.D.,D’Abreu,A.,Burock,J.,Curtin,A.,Wu,C.K., Morris,J.C.,2005.Clinicianassessmentofthedrivingcompetenceofpatients withdementia.JournaloftheAmericanGeriatricsSociety53(5),829–833. Ott,B.R.,Festa,E.K.,Amick,M.M.,Grace,J.,Davis,J.D.,Heindel,W.C.,2008.

Com-puterizedmazenavigationandon-roadperformancebydriverswithdementia. JournalofGeriatricPsychiatryandNeurology21(1),18–25.

Parasuraman, R.,Nestor,P.G.,1991.Attentionand drivingskillsin agingand Alzheimer’sdisease.HumanFactors33(5),539–557.

Pendlebury,S.T.,Mariz,J.,Bull,L.,Mehta,Z.,Rothwell,P.M.,2012.MoCA ACE-R, andMMSE versustheNational InstituteofNeurological Disordersand Stroke-CanadianStrokeNetworkVascularCognitiveImpairment Harmoniza-tionStandardsNeuropsychologicalBatteryafterTIAandstroke.Stroke43(2), 464–469.

Radford,K.,2001.ValidationoftheStrokeDriversScreeningAssessmentforpatients withanacquiredneurologicaldisability.PhDthesis.UniversityofNottingham, Nottingham.

Rebok,G.W.,Keyl,P.M.,Bylsma,F.W.,Blaustein,M.J.,Tune,L.,1994.Theeffectsof Alzheimerdiseaseondriving-relatedabilities.AlzheimerDiseaseandAssociated Disorders8(4),228–240.

Reger,M.,Welsh,R.,Watson,G.,Cholerton,B.,Baker,L.,Craft,S.,2004.The relation-shipbetweenneuropsychologicalfunctioninganddrivingabilityindementia: ameta-analysis.Neuropsychology18(1),85–93.

Reyes,M.A.,Lloret,S.P.,Gerscovich,E.R.,Martin,M.E.,Leiguarda,R.,Merello,M., 2009.Addenbrooke’sCognitiveExaminationvalidationinParkinson’sdisease. EuropeanJournalofNeurology16(1),142–147.

Rizzo,M.,2011.Impaireddrivingfrommedicalconditions:a70-year-oldmantrying todecideifheshouldcontinuedriving.TheJournaloftheAmericanMedical305 (10),1018–1026.

Rizzo,M.,Kellison,I.L.,2010.Thebrainontheroad.In:Marcotte,T.D.,Grant,I. (Eds.),NeuropsychologyofEverydayFunctioning.GuilfordPress,NewYork,pp. 168–207.

Sadek, J.R., van Gorp, W.G., 2010. Prediction of vocational functioning from neuropsychologicalperformance.In:Marcotte,T.D.,Grant,I.(Eds.), Neuropsy-chologyofEverydayFunctioning.GuilfordPress,NewYork,pp.113–135. Selander,H.,Lee,H.C.,Johansson,K.,Falkmer,T.,2011.Olderdrivers:on-roadand

off-roadtestresults.AccidentAnalysisandPrevention43(4),1348–1354. Simões,M.R.,Firmino,H.,Sousa,L.B.,Pinho,M.S.,2011.AddenbrookeCognitive

ExaminationRevised(ACE-R):Portugueseadaptation,validationand norm-ing.In:39thCongressofEuropeanAssociationofGeriatricPsychiatry,Porto, Portugal.

Strauss,A.,Sherman,E.,Spreen,O.,2006.CompendiumofNeuropsychologicalTests: Administration,NormsandCommentary,3rded.OxfordUniversityPress,New York.

Torralva,T.,Roca,M.,Gleichgerrcht,E.,Bonifacio,A.,Raimondi,C.,Manes,F.,2011. Validacióndelaversiónenespa ˜noldelAddenbrooke’sCognitive Examination-Revisado(ACE-R).Neurología26(6),351–356.

Uc,E.Y.,Rizzo,M.,Anderson,S.W.,Shi,Q.,Dawson,J.D.,2004.Driverroute-following andsafetyerrorsinearlyAlzheimerdisease.Neurology63(5),832–837. Whelihan,W.M.,DiCarlo,M.A.,Paul,R.,2005.Therelationshipofneuropsychological

functioningtodrivingcompetenceinolderpersonswithearlycognitivedecline. ArchivesofClinicalNeuropsychology20(2),217–228.

Wolfe,P.L.,Clark,J.A.,2012.Drivingcapacity.In:Demakis,G.D.(Ed.),Civil capaci-tiesinClinicalNeuropsychology:ResearchFindingsandPracticalApplications. OxfordUniversityPress,NewYork,pp.121–138.

Yoshida,H.,Terada,S.,Honda,H.,Kishimoto,Y.,Takeda,N.,Oshima,E.,Hirayama, K.,Yokota,O.,Uchitomi,Y.,2012.ValidationoftherevisedAddenbrooke’s CognitiveExamination(ACE-R)fordetectingmildcognitiveimpairmentand dementia in a Japanese population. International Psychogeriatrics24 (1), 28–37.