9
This article was downloaded by: [University of Windsor] On: 16 November 2014, At: 02:18 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK The Clinical Neuropsychologist Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ntcn20 Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To- Drive Assessment of Older Adults Rudi De Raedt & Ingrid Ponjaert-Kristoffersen Published online: 09 Aug 2010. To cite this article: Rudi De Raedt & Ingrid Ponjaert-Kristoffersen (2001) Short Cognitive/ Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults, The Clinical Neuropsychologist, 15:3, 329-336 To link to this article: http://dx.doi.org/10.1076/clin.15.3.329.10277 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms- and-conditions

Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

  • Upload
    ingrid

  • View
    213

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

This article was downloaded by: [University of Windsor]On: 16 November 2014, At: 02:18Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

The Clinical NeuropsychologistPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/ntcn20

Short Cognitive/NeuropsychologicalTest Battery for First-Tier Fitness-To-Drive Assessment of Older AdultsRudi De Raedt & Ingrid Ponjaert-KristoffersenPublished online: 09 Aug 2010.

To cite this article: Rudi De Raedt & Ingrid Ponjaert-Kristoffersen (2001) Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults, TheClinical Neuropsychologist, 15:3, 329-336

To link to this article: http://dx.doi.org/10.1076/clin.15.3.329.10277

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Page 2: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

Rudi De Raedt and Ingrid Ponjaert-KristoffersenDepartment of Developmental and Lifespan Psychology, Free University of Brussels, Brussel, Belgium

ABSTRACT

Since ®tness-to-drive evaluation of elderly drivers has become an important issue, we developed a short ®rst-tier screening battery to evaluate the necessity for further referral to specialised centres. Our sampleconsisted of 84 subjects between 65 and 96 years who came to the Belgian Road Safety institute for a ®tness-to-drive evaluation. Using cross-validated discriminant analyses, the predictive power of a battery consistingof the Trail Making Test, Part A, a visual acuity test, a clock drawing test, the Mini-Mental State Examination(MMSE) and age was analysed. The judgement by an independent driver instructor (®t-to-drive vs. notunconditional ®t-to-drive), based on a real world road test was used as the dependent variable. Classi®cationfunctions based on the signi®cant discriminant function yielded a speci®city score of 85% (subjects ®t-to-drive correctly classi®ed) and a sensitivity score of 80% (subjects as not unconditional ®t-to-drive correctlyclassi®ed). These results highlight the potential value of a short screening instrument that can be used inprimary health care settings. This instrument may be useful as a ®rst step in a multi-tier assessmentprocedure.

INTRODUCTION

One of the new scienti®c and social challenges

concerns the older car driver. Indeed, seniors

constitute a very rapidly increasing age group in

the population of car drivers (Hakamies-Blom-

qvist, 1996) and maximum mobility is a very

important issue for this age group. Compared to

the whole population, older drivers are a low-risk

group concerning accidents. However, taking into

account the reductions in distances driven with

increasing age, their risk per driven mile becomes

much higher (Evans, 1988). On the other hand,

compensatory mechanisms might be useful to

prevent accidents (Hakamies-Blomqvist, 1993),

although many older people may not use these

protective strategies (Christ, 1996). Therefore, an

instrument to detect drivers with problems might

enable preventive measures such as rehabilitation

programs focussing on the use of compensation

strategies. This is important since it has been

demonstrated that driving cessation leads to

increases in depressive symptoms (Marottoli

et al., 1997). Therefore, well-founded advice

concerning driving ability is very important for

elderly people.

Since primary care practitioners are often

confronted with elderly people suffering cogni-

tive problems, it may be important for them to

obtain some knowledge concerning the risk asso-

ciated with car driving. However, since ®tness-to-

drive evaluation is a complex matter that requires

a multidisciplinary approach, it is not sensible

that primary care practitioners would have to

make ®nal decisions concerning the driving abil-

ity of their older patients. In many countries,

Address correspondence to: Rudi De Raedt, Department of Developmental and Lifespan Psychology, FreeUniversity of Brussels, Pleinlaan 2 (3C247), B-1050 Brussel, Belgium. Tel.: � 32 2 629 36 22. Fax: � 32 2 629 2532. E-mail: [email protected] for publication: April 11, 2001.

The Clinical Neuropsychologist 1385-4046/01/1503-329$16.002001, Vol. 15, No. 3, pp. 329±336 # Swets & Zeitlinger

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 3: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

specialised centres address the ®tness-to-drive

issue, often using a road test in order to observe

how the person behaves in real world traf®c

situations. Primary care practitioners refer more

and more elderly people to such ®tness-to-drive

assessment centres, though it is not always easy to

assess the necessity for referral. Moreover, former

research revealed that common medical examina-

tion could not distinguish between safe and

unsafe drivers while cognitive decline was

revealed to be more related with driving problems

(Johansson et al., 1996). A common problem is

that driving-related cognitive impairment is dif®-

cult to estimate in a primary health care setting.

Therefore, we developed a short test battery con-

sisting mainly of cognitive tests that can be used

as a ®rst indicator of driving-related problems.

The test is not aimed as a conclusive instrument,

but as a ®rst indicator for further referral.

The aim of this paper is to identify the pre-

dictive power of our short neuropsychological test

battery in relation to the judgement of expert

driving instructors in an of®cial ®tness-to-drive

assessment centre. The choice of the tests was

based on a survey of the literature concerning car

driving and cognitive functioning. A (mainly)

cognitive approach has been adopted. From a

cognitive / neuropsychological viewpoint, car dri-

ving can be considered as a very complex activity.

However, most of the time, its complexity is not

experienced because many aspects of driving are

highly automated. When a cognitive problem

arises, however, a distortion of the complex

information processing system can produce very

dangerous situations. This may be the reason why

several researchers could demonstrate that cogni-

tive variables are an important causal factor in

crashes of older drivers (Lundberg, Hakamies-

Blomqvist, Almkvist, & Johansson, 1998). We

used models of driver behaviour and cognitive

ageing to identify the cognitive skills relevant to

safe car driving and to draw a picture of the

in¯uence of aging on the cognitive information

processing system with respect to car driving (De

Raedt, 2000). All these functions were operatio-

nalised by selected neuropsychological tests. The

visuo-sensory, visuo-perceptual, and visuo-spatial

functions, the different basic attention functions,

the useful ®eld of view, automatic versus con-

trolled processes, cognitive ¯exibility, the psy-

chomotor system, and executive planning

functions were analysed in depth. Two instru-

ments were constructed, a computer-based battery

for detailed neuropsychological assessment (De

Raedt & Kristoffersen, 2000a) and the screening

battery presented in this paper. An important

criterion for a subtest to be included in the short

screening battery was that it had to be short, easy

to administer in a primary health care setting, and

with the least discomfort for the older patient.

Other researchers already highlighted that some

easy to administer tests could be valid instruments

for such a ®tness-to-drive screening procedure.

For example, Marottoli and co-workers demon-

strated the relevance of visual acuity (Marottoli

et al., 1998). Much other research revealed the

relevance of a drawing task and the Mini-Mental

State Examination (MMSE) (Johansson et al.,

1996) and the relation of the Trail Making

Test, Part A and adverse driving events (Stutts,

Stewart, & Martell, 1998). The purpose of this

research is not to investigate the predictive power

of these separate measures again, but to develop

a combined short screening battery based on

these tests with an easy to calculate single score

and to analyse the discriminatory power of this

unique new measure. In contrast to other re-

search projects (Trobe et al., 1996), our study

focuses on elderly people without diagnosis of

dementia.

METHOD

The ethics committee of the medical faculty of the`Free University of Brussels' approved the researchprotocol.

Subjects and Study SiteThe sample included 84 car drivers (24 women and60 men) from 65 to 96 years old (M � 78.6,SD � 6.8). They were referred to the CARA (Fit-ness-to-drive evaluation centre) department of theBelgian Road Safety Institute for a ®tness-to-driveevaluation, and they volunteered participation in thestudy after informed consent (two referred subjectsdid not collaborate). The participants were referredby their physician or directly by their insurancecompany.

330 RUDI DE RAEDT AND INGRID PONJAERT-KRISTOFFERSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 4: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

Only healthy people without diagnosed neurolo-gical pathologies that might interfere with neuro-psychological functioning were included. Moreover,people who were suspected of dementia / cognitivedecline by their family doctor were excluded. Toevaluate medical condition, a comprehensive ques-tionnaire had to be completed by the family doctor.When, based on the questionnaire, the doctor ofCARA was uncertain if the subject met the inclusioncriteria of the study, she examined the subjectherself. Subjects in the sample had to meet theBelgian ®tness-to-drive criterion concerning visualacuity. Visual acuity in the central ®eld of view (far-sight; with optical correction) had to reach 4 /10 atthe moment the research was conducted, but the newcriterion is currently 5 /10. Since many people werereferred because of a change of insurance companyor because of minor accidents, our research sampleconsisted of a well-balanced group of subjects withand without driving problems.

Items Included in the BatteryIn one of the statistical analyses, the dependentvariable is one combined total score for the wholebattery: In order to obtain equal weights for eachitem of the battery, all scores were calculated to a600-point scale.

The `Ergovision' testing device (Essilor: 1, RueThomas Edison, 94028 CreÂteil Cedex France) wasused to assess static visual acuity (far-sight/central),on a scale from 0 to 10. It is a regular acuity test inwhich digits of different sizes have to be perceivedcorrectly. Each score was multiplied by 60.

The Trail Making Test, Part A (Reitan, 1955)was used to assess selective attention with visualscanning and search. During this task, a paper withrandom distributed circled numbers is presented.Subjects are asked to draw lines between thenumbers in the correct order. The time needed (inseconds) to complete the test was subtracted from200 and multiplied by three. (Times greater than200 s were scored as 0).

The Mini-Mental Status Examination (MMSE)(Folstein, Folstein, & Mc Hugh, 1975) was trans-lated into Dutch and French by the authors and usedto assess global cognitive functioning. The totalscore (max 30) was multiplied by 20. During thistest, `orientation in time and space', `memoryencoding', `attention', `memory retrieval', `lan-guage' and `visuo-spatial function' are brie¯yevaluated using simple questions and simple tasks.

A clock drawing test (Spreen & Strauss, 1998)was administered to assess visuo-spatial function.People were asked to draw a clock with decimaldigits indicating 11.10 h. Four points were scored:

Clock face complete (max 10% missing); All digitspresent; Axes 12-6 and 9-3 must be placed at rightangles (max 20 difference); hands correctly placed(they must be placed closer to 11 and 2 than to theneighbouring digits). The total score (max � 4) wasmultiplied by 150.

Finally, the age of each person was subtractedfrom 105 and multiplied by 15 (the minimum age inthe population to assess is 65 years).

The total maximum score on the whole battery(5� 600 � 3000) was divided by 100 to yield atransformed maximum score of 30.

Road TestA road test was administered by two driver instruc-tors of the ®tness-to-drive assessment centre(CARA), using a dual-brake car. The driver instruc-tors were blind to the predictive test results of thesubjects. The road test was a standardised in-traf®c35 km trajectory. A part of the trajectory was in townareas and another part outside of town areas. Differ-ent traf®c situations enabled relevant observations.As an example, participants were required to makecomplex left turns merging onto a main road. Parti-cipants encountered also many other `challenge'situations such as joining the traf®c stream and lanechanging on a crowded highway. Afterwards, theexaminer ®lled in a detailed evaluation grid: TheTest Ride for Investigating Practical ®tness-to-drive /Belgian version (TRIP). The TRIP protocol wasdeveloped by Brouwer and co-workers at the neu-ropsychology/gerontology department of the Uni-versity of Groningen (The Netherlands) in acollaboration with CBR (®tness-to-drive assessmentCentre of The Netherlands) and CARA (®tness-to-driveassessment centre of Belgium). The instrument(Belgian version) consists of 11 dimensions, eachrated on 3- and 4-point scales. At the end, theinstructors were asked if, based on their observations(qualitative approach), they considered the subjectas unconditional ®t-to-drive (without restrictions),with restrictions (conditional ®t-to-drive), or as not®t-to-drive. Since the aim of our analyses was toisolate subjects in need of further evaluation, twocategories were used as the dependent variable:`Unconditional ®t-to-drive' and `not unconditional®t-to-drive'.

In the latter group, 10 subjects were judged un®t-to-drive, 30 ®t-to-drive with restrictions (condi-tional). A total of 44 subjects were unconditional ®t-to-drive. A total of 38 road tests had been videotaped(video-equipment in the back of the car) and werescored by another independent driver instructorblinded to the condition and the predictive testresults of the subjects.

FITNESS-TO-DRIVE ASSESSMENT 331

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 5: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

Statistical AnalysesDiscriminant function analyses were used to analysethe predictive value of our test battery to the judge-ment of an expert driver instructor. The analyseswere started with a stepwise discriminant functionanalysis to highlight the relative contribution of thevarious measures employed to the ®nal classi®ca-tion. The raw test scores of each test were consideredas independent variables with the judgement of theinstructors (drivers judged unconditional `®t-to-drive' vs. drivers judged not unconditional `®t-to-drive') as dependent variables (F-to enter was setat 2.5).

Because unequal groups as de®ned by thedependent variable (®t / un®t) can in¯uence thediscriminant function and the classi®cation ofsubjects, we selected random subjects from the poolof the largest group in order to obtain dependentvariable groups of equal size. A total of 80 subjectsremained in the analysis group of which 40 werejudged as unconditional ®t-to-drive and 40 as notunconditional ®t-to-drive. Based on the signi®cantdiscriminant function, a classi®cation function wascalculated to classify subjects into the two categories(®t/un®t). In this way, the percentage of subjectscorrectly classi®ed by the model was calculated.

Since it is also important to demonstrate realworld usefulness of the instrument with one singleeasy to calculate score, another discriminant func-tion analysis has been performed with only onevariable (the total score (0±30)) entered as indepen-dent variable. Entering the test scores step by stepdid not yield classi®cation functions based on whichone single cut-off score can be identi®ed. In order tocross-validate the results of the latter analysis, thediscriminant function/classi®cation function wascalculated on a random-selected half of the sample(on 40 random-selected subjects) proportional to thegroups (®t / un®t) and applied to the other half of thesample (the other 40 subjects).

Interrater reliability for the road test judgementwas evaluated by calculating the percentage ofidentical judgements between the two driver instruc-tors (see section on road test). Three categories wereused: unconditional ®t-to-drive, un®t-to-drive, ®t-to-drive on condition.

All the analyses were perfomated using STATIS-TICA 5.1. for Windows.

RESULTS

The stepwise discriminant function analysis with

the raw test scores of each test considered as

independent variables and the judgement of the

driving instructors as the dependent variable

(®t /un®t) yielded a signi®cant discriminant func-

tion (Wilks' Lambda � 0.60; F(4, 75) � 12.5,

p< 0.0001). The speci®city score (percentage

subjects ®t-to-drive correctly classi®ed) was

87.5%. The sensitivity (percentage subjects as not

unconditional ®t-to-drive correctly classi®ed) was

75%. This yields an overall hit ratio of 81.25%. The

MMSE score was not selected by the model since

this variable did not add signi®cant discriminatory

power (F � 2, 5) (The discriminant function ana-

lysis summary is outlined on Table 1).

The discriminant function analysis with the

combined test score as the independent variable

(one variable) and the judgement of the driving

instructors as the dependent variable (®t / un®t)

yielded a signi®cant discriminant function

(Wilks' Lambda � 0.63; F(1, 38) � 22.6, p <0.0001). The speci®city score (percentage sub-

jects ®t-to-drive correctly classi®ed) was 85%.

The sensitivity (percentage subjects as not uncon-

ditional ®t-to-drive correctly classi®ed) was 80%.

This yielded an overall hit ratio of 82.5%. These

ratios coincide with a cut-off score of 24 / 30

(people should be referred below score 24).

Concerning interrater reliability, 79% of the

judgements of the two independent driver instruc-

tors were identical (unconditional ®t-to-drive,

un®t-to-drive, ®t-to-drive on condition).

DISCUSSION

Our combined measure of visuo-sensory function,

selective attention with visual scanning and

search, visuo-spatial function, general cognitive

functioning, and age yielded high sensitivity and

Table 1. Discriminant Function Analysis SummaryWith the Judgement of the Driver Instructorsas Dependent Variables and the Untrans-formed Test scores as Independent Variables.

Wilk's F to p-levelLambda remove

Trail Making Test, Part A (s) 0.67 8.4 0.005Age (years) 0.64 4.8 0.03Clock drawing (0±4) 0.62 3.0 0.09Visual acuity (0±10) 0.62 2.7 0.11

332 RUDI DE RAEDT AND INGRID PONJAERT-KRISTOFFERSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 6: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

speci®city values with respect to the judgement of

a professional driver instructor. This highlights

the potential usefulness of this instrument, which

takes approximately 25 min to administer and

score. The correlations of the Trail Making Test,

Part A and the MMSE with driving performance

have already been observed in many former

research projects on older drivers (Odenheimer

et al., 1994). Moreover, the reliability of the Trail

Making Test, Part A is well documented, with

reported retest reliability coef®cients mostly above

.60 but several above .80 and .90 (Lezak, 1995). In a

study by Snow and co-workers (Snow, Tierney,

Zorzitto, Fisher, & Reid, 1998), the 1-year retest

reliability for the Trail Making Test, Part A in 100

elderly subjects was .64. However, this moderate

result may be explained by the fact that, in a

population of older subjects, deterioration within

a 1-year period is possible. For the MMSE, satis-

factory test-retest reliability has also been demon-

strated (Lezak, 1995), ranging between .63 and .83

(24-hr test±retest reliability with different exami-

ners). Marottoli and co-workers (1994) demon-

strated that the copying subtask of the MMSE

(also a drawing task) predicted adverse traf®c

events. In our population, we observed a ceiling

effect on that copying task. Therefore, we used a

clock drawing test. This test is widely used with

elderly persons and yields high retest reliability

scores ranging between .78 and .97. (Spreen &

Strauss, 1998). Concerning visual acuity, mixed

results are found in the literature. Although, some

studies report no relation with safe driving (Johans-

son et al., 1996), others do (Hills, 1980). However,

we added visual acuity to the battery because it is

the only visuo-sensory measure that can be admi-

nistered easily in every primary health care setting.

The correlation of age (in the 65� age group) with

safe driving is a consistent ®nding (Odenheimer

et al., 1994). Therefore, this easy to grasp variable

was included in the battery. However, it has to be

stated that age alone might never be a reason to

recommend driving cessation, since the variability

of cognitive ability is very high in this age group

(Welford, 1993).

The interrater reliability of the judgement of

the driver instructor was revealed to be high.

In our sample, we observed that only four out

of ®ve items were selected to be included in the

stepwise analysis. The MMSE score did not add

enough extra variance to be included. The item

that showed the strongest discriminatory power

was the Trail Making Test, Part A. The absence of

predictive power for the MMSE score is in line

with research by Bieliauskas and co-workers

(Bieliauskas, Roper, Trobe, Green, & Lacy,

1998), in which patients with dementia were

evaluated on an on-the-road driving test. On the

other hand, in a study by Fitten and co-workers

(1995) on dementia and driving, the MMSE score

was among the variables that correlated best with

driving abilities. However, it has to be highlighted

that the sample sizes were very small in both

studies. In a large scale study with Alzheimer

patients by Trobe et al. (1996), the MMSE score

did not predict future crashes or violations. In our

study, it is not surprising that the MMSE scores

did not show much discriminatory power, since

the MMSE scores for both groups were rather

high (27.8 for the ®t-to-drive group and 25.6 for

the un®t-to-drive group). However, as suggested

by the results of a study by Odenheimer et al.

(1994), it might be that the MMSE is more relevant

in mixed populations. In this research program,

subjects with a wide range of cognitive ability were

included, and the correlation between an in-traf®c

score and the MMSE score was .72. The sample of

this research is more representative of the general

population of older adults. Our sample includes

only subjects without dementia, and the studies of

Bieliauskas et al. (1998), Fitten et al. (1995) and

Trobe et al. (1996) include only subjects with

dementia. Therefore, the MMSE score might still

be of importance in a screening battery to be used in

primary health care settings. This assumption is, of

course, in need of further investigation.

The overall hit rate of the battery with the single

global score and after cross-validation was 82.5%.

The results of the stepwise analyses without the

MMSE score on the whole population yielded an

overall hit score of 81.5%. These results are indeed

indicative for the fact that, in our population, the

MMSE score is not very important.

When the battery is used to isolate subjects in

need of further evaluation, all subjects with a

global score lower than the cut-off score (24 /

30) should be referred. In this way, 15% of good

drivers would be referred unnecessarily, while

FITNESS-TO-DRIVE ASSESSMENT 333

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 7: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

20% of drivers with problems will be missed. The

fact that these results are cross-validated is indi-

cative of `real world' predictive power. A total of

50% of the subjects used for the analyses were

considered as not unconditional ®t-to-drive,

which is much more than would be expected in

the overall population, but oversampling may be

used for validity research since a rare event is

statistically dif®cult to predict. Therefore, over-

sampling is a commonly used research strategy in

this ®eld (Goode et al., 1998). Moreover, because

many subjects were referred because of an insur-

ance company change or because of minor acci-

dents, our sample re¯ects a good balance between

drivers with and without problems. However, it

has to be highlighted once more that our popula-

tion may not be entirely representative for a group

that would present to primary health care. There-

fore, this battery needs re-testing in a broader

population. Moreover, further research is needed

to uncover whether the use of these tests, which

can be administered to cognitively impaired peo-

ple (Lezak, 1995), is also predictive of impair-

ment in a population suffering from dementia.

The importance of identifying problem drivers

in a primary health care setting has already been

highlighted in several publications (O'Neill,

1992). It has been suggested that it might be easy

and less time consuming to refer all subjects with

an MMSE score< 30. When applied on the same

random cross-validation sample used for our

analyses, referring all subjects with an MMSE

score < 30 yields indeed a very high sensitivity

score of 95%, but a speci®city score of only 15%.

This corresponds with a total hit score of 55%,

which is only slightly better than chance. Our

battery yielded a total hit score of 82.5%, high-

lighting the added value of our composed battery

over the use of a very simple and straightforward

measure.

Janke and Eberhard (1998) recently proposed a

three-tier assessment system. The authors high-

lighted the importance of a ®rst brief and inex-

pensive screening test to ¯ag drivers not reported

to the agency. They also proposed a short ®rst-tier

screening test. Using a logistic regression proce-

dure based on a population of 31 volunteers and

65 referrals, they analysed the predictive power of

their instrument to predict referrals versus volun-

teers. The best discriminating variables entered in

the model included a computerised Trail Making

Test, Part A, a contrast sensitivity test, and the

number of problems observed while testing (such

as tremor; dif®culty in understanding). The clas-

si®cation function they applied in their model

yielded a high speci®city score of 97% (volun-

teers classi®ed correctly), but with a sensitivity of

only 63% (referrals classi®ed correctly). The

classi®cation function of our discriminant model

(we chose an optimum between sensitivity and

speci®city) yielded a speci®city of 85% (drivers

judged ®t-to-drive classi®ed correctly) and a sen-

sitivity of 80% (drivers judged as not uncondi-

tional ®t-to-drive classi®ed correctly). As

mentioned by Janke and Eberhard (1998), it has

to be noted that the referral status used in their

study acts only as a ¯awed surrogate of being

suf®ciently impaired to warrant further evalua-

tion. Therefore, we used, as the dependent vari-

able, real world driving performance as judged by

experts, as opposed to referral status. Moreover

our groups, as de®ned by grouping variables, were

equalised to reduce chance effects and we cross-

validated our results. Another strength of our

study was the absence of volunteer bias, since

nearly the entire referred subjects group partici-

pated in our study (only 2 out of 84 did not).

When reviewing the literature, mixed ®ndings are

reported concerning the possibilities of a screen-

ing measure for driving safety. Our population

was quite homogeneous, with smaller standard

deviations than in mixed populations or popula-

tions with dementia. The fact that we used a ®xed

route with many dif®cult `challenge' situations in

an attempt to standardise the test situation and to

enable many different relevant observations may

also be an important reason for the high prediction

accuracy we obtained. Withaar and co-workers

(1997) used the same observation grid that we

applied (the TRIP), but their participants drove

their own car in their own neighbourhood, en-

countering less `challenge' situations to enhance

the ecological validity of the observations. As

Bieliauskas et al. (1998) postulate, neuropsycho-

logical measures may be more predictive in chal-

lenge-related situations. Indeed, in routine

situations, automatic procedural routines might

be suf®cient, while cognitive neuropsychological

334 RUDI DE RAEDT AND INGRID PONJAERT-KRISTOFFERSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 8: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

abilities may be required in challenging and

unpredictable situations, in which controlled stra-

tegies come into play.

The results of our study provide support for the

possibility of a short ®rst-tier neuropsychological

screening test to distinguish elderly people in

need of further evaluation. Second-tier neuro-

psychological evaluation has to focus on detailed

analyses of neuropsychological functions and

their relation to driving performance (De Raedt

& Kristoffersen, 2000a) and in-depth analyses

concerning the relationship between cognitive

functioning and accidents may be very important

(De Raedt & Kristoffersen, in press). In this way,

driver strengths and weaknesses can be discov-

ered, highlighting possibilities for compensation

and rehabilitation. We believe that society has the

responsibility to tackle the problem of older

drivers in order to create opportunities for their

maximum mobility in safe conditions. Research

must not focus solely on shortcomings but also on

the possibilities for older people to adapt to these

shortcomings.

Therefore, every older driver should be given

the opportunity to be evaluated using a road test

(eventually during a third-tier assessment) to

uncover how she / he behaves and adapts in real

world traf®c situations. Such a driving test might

be indicative for tactical compensation strategies

like adapted speed choice, adapted distance to the

car in front, and compensatory anticipation beha-

viour. However, empirical research concerning

the positive effects of compensation strategies is

almost lacking at the moment. Nevertheless, we

could demonstrate that compensation strategies

can be successful in avoiding accidents (De Raedt

& Kristoffersen, 2000b). On the other hand, for a

part of the older population, it may become

obvious that safe traf®c participation as a car

driver is no longer possible. In these cases, atten-

tion should be focused on coping with driving

cessation (O'Neill, 1997).

ACKNOWLEDGEMENT

We would like to thank the staff of the CARA

department for their help in completing this study.

We also wish to thank Wiebo Brouwer and

Frederiec Withaar (neuropsychology-gerontology

department/groningen University, The Nether-

lands) for their advice. The authors owe special

thanks to the Belgian Road Safety Institute and to

the Research Council of the Free University

Brussels who funded this research project.

REFERENCES

Bieliauskas, L.A., Roper, B.R., Trobe, J., Green, P., &Lacy, M. (1998). Cognitive measures, drivingsafety, and Alzheimer's disease. The ClinicalNeuropsychologist, 12, 206±212.

Christ, R. (1996). Ageing and driving ± decreasingmental and physical abilities and increasing com-pensatory abilities. IATSS Research , 20, 43±52.

De Raedt, R. (2000). Cognitive / neuropsychologicalfunctioning and compensation related to car drivingperformance in older adults (doctoral dissertation).Brussels: Vrije Universiteit Brussel.

De Raedt, R., & Ponjaert-Kristoffersen, I. (2000a). Therelationship between cognitive / neuropsychologicalfactors and car driving performance in older adults.Journal of the American Geriatrics Soceity, 48,1664±1668.

De Raedt, R., & Ponjaert-Kristoffersen, I. (in press).Predicting at-fault car accidents of older drivers.Accident Analysis and Prevention.

De Raedt, R., & Ponjaert-Kristoffersen, I. (2000b). Canstrategic and tactical compensation reduce crashrisk in older drivers. Age and Ageing, 29, 517±521.

Evans, L. (1988). Older driver involvement in fatal andsevere traf®c crashes. Journal of Gerontology:Social Sciences, 43, 186±193.

Fitten, L.J., Perryman, K.M., Wilkinson, C.J., Little,R.J., Burns, M.M., Pachana, M., Mervis, J.R.,Malmgren, R., Siembieda, D.W., & Ganzell, S.(1995). Alzheimer and vascular dementias anddriving: A prospective road and laboratory study.Journal of the American Medical Association, 273,1360±1365.

Folstein, M.L., Folstein, S.E., & Mc Hugh, P.R. (1975).Mini-Mental State: A practical method for gradingthe cognitive status of patients for the clinician.Journal of Psychiatric Research, 12, 189±198.

Goode, K.T., Ball, K.K., Sloane, M., Roenker, D.L.,Roth, D.L., Myers, R.S., & Owsley, C. (1998).Useful ®eld of view and other neurocognitiveindicators of crash risk in older adults. Journal ofClinical Psychology in Medical Settings, 5, 425±440.

Hair, J.F. (1998). Multivariate data analysis. UpperSaddle River: Prentice-Hall.

Hakamies-Blomqvist, L. (1993). Compensation inolder drivers as re¯ected in their fatal accidents.Accident Analysis and Prevention, 26, 107±112.

FITNESS-TO-DRIVE ASSESSMENT 335

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14

Page 9: Short Cognitive/Neuropsychological Test Battery for First-Tier Fitness-To-Drive Assessment of Older Adults

Hakamies-Blomqvist, L. (1996). Research on olderdrivers: A review. IATSS Research, 20, 91±101.

Hills, B.L. (1980). Vision, visibility and perception indriving. Perception, 9, 183±16.

Janke, M.K., & Eberhard, J.W. (1998). Assessingmedically impaired older drivers in a licensingagency setting. Accident Analysis and Prevention,30, 347±361.

Johansson, K., Bronge, L., Lundberg, C., Persson, A.,Seideman, M., & Viitanen, M. (1996). Can aphysician recognize an older driver with increasedcrash risk potential? Journal of the AmericanGeriatrics Society, 44, 1198±1204.

Lezak, M.D. (1995). Neuropsychological assessment(3rd ed.). New York: Oxford University Press.

Lundberg, C., Hakamies-Blomqvist, L., Almkvist, O., &Johansson, K. (1998). Impairments of some cognitivefunctions are common in crash-involved older dri-vers. Accident Analysis and Prevention, 30, 371±377.

Marottoli, R.A., Cooney, L.M., Wagner, D.R., Doucette,J., & Tinetti, M.E. (1994). Predictors of automobilecrashes and moving violations among elderly drivers.Annals of Internal Medicine, 121, 842±846.

Marottoli, R.A., Mendes De Leon, C.F., Glass, T.A.,Williams, C.S., Cooney, L.M., Berkman, L.F., &Tinetti, M.E. (1997). Driving cessation andincreased depressive symptoms: Prospective evi-dence from the New Haven EPESE. Journal of theAmerican Geriatrics Society, 45, 202±206.

Marottoli, R.A., Richardson, E.D., Stowe, M.H., Miller,E.G., Brass, L.M., Cooney, L.M., Jr., & Tinetti, M.E.(1998). Development of a test battery to identifyolder drivers at risk for self-reported adverse drivingevents. Journal of the American Geriatrics Society,30, 562±568.

Odenheimer, G.L., Beaudet, M., Jette, A.M., Albert,M.S., Grande, L., & Minaker, K.L. (1994). Perfor-

mance-based driving evaluation of the elderlydriver: Safety, reliability, and validity. Journal ofGerontology, 49, M153±M159.

O'Neill. (1992). Physicians, elderly drivers, anddementia. Lancet, 339, 41±43.

O'Neill, D. (1997). Predicting and coping with theconsequences of stopping driving. Alzheimer Dis-ease and Associated Disorders, 11, 70±72.

Reitan, R.M. (1955). The relation of the Trail MakingTest to organic brain damage. Journal of ConsultingPsychology, 19, 393±394.

Snow, W.G., Tierney, M.C., Zorzitto, M.L., Fisher,R.H., & Reid, D.W. (1998). One-year test±retestreliability of selected neuropsychological tests inolder adults. Journal of Clinical and ExperimentalNeuropsychology, 10, 60.

Spreen, O., & Strauss, E. (1998). A compendium ofneuropychological tests: Administration, norms, andcommentary. New York: Oxford University Press.

Stutts, J.C., Stewart, J.R., & Martell, C. (1998).Cognitive test performance and crash risk in anolder driver population. Accident Analysis andPrevention, 30, 337±46.

Trobe, J.D., Waller, P.F., Cook-Flannagan, C.A.Teshima, S.M., & Bieliauskas, L.A. (1996). Crashesand violations among drivers with AlzheimerDisease. Archives of Neurology, 53, 411±416.

Welford, A.T. (1993). The gerontological balancesheet. In J. Cerella, J. Rybash, W. Hoyer, & M.L.Commons (Eds.), Adult information processing:Limits on loss. San Diego: Academic Press.

Withaar, F., Brouwer, W., van Zomeren, E., & Deelman,B. (1997). Predicting performance of instrumentalactivities of daily living in neuropsychologi-cally impaired older adults. Journal of the Interna-tional Neuropsychological Society (abstract), 3,215.

336 RUDI DE RAEDT AND INGRID PONJAERT-KRISTOFFERSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f W

inds

or]

at 0

2:18

16

Nov

embe

r 20

14