European Geriatric Medicine 4 (2013) 154–160

Research paper

Older drivers with cognitive impairment: Perceived changes in driving skills,driving-related discomfort and self-regulation of driving

A. Meng a,b,*, A. Siren b, T.W. Teasdale a

a Department of Psychology, University of Copenhagen, Øster Farimagsgade 2a, 1353 Copenhagen, Denmarkb Department of Transport, Technical University of Denmark, Bygningstorvet 115, 2800 Kgs, Lyngby, Denmark

A R T I C L E I N F O

Article history:

Received 5 November 2012

Accepted 7 January 2013

Available online 29 January 2013

Keywords:

Cognition

Compensatory behaviour

Driving-related stress

Self-monitoring of driving

Traffic safety

A B S T R A C T

The results of a previous study indicate that in general, older drivers who recognise cognitive problems

show realistic self-assessment of changes in their driving skills and that driving-related discomfort may

function as an indirect monitoring of driving ability, contributing to their safe driving performance. The

aim of the present study was to examine whether these findings also apply to cognitively impaired older

drivers. Structured face-to-face interviews were conducted with 25 cognitively impaired older drivers.

The results showed that the participants were most likely to report their driving skills as unchanged.

There was an association between level of discomfort and avoidance of driving situations, but not

between cognitive status and discomfort or avoidance. The results suggest that cognitively impaired

older drivers constitute a unique group; while cognitively impaired older drivers may recognise

cognitive problems, they tend not to recognise changes to their driving, which may reflect reluctance to

acknowledge the impact of cognitive impairment on their driving. Furthermore, the results suggest that

driving-related discomfort plays an important role in the self-regulation of driving among cognitively

impaired older drivers. However, it is less clear what triggers driving-related discomfort among

cognitively impaired older drivers indicating that it may be a less reliable aspect of their self-monitoring

of driving ability.

� 2013 Elsevier Masson SAS and European Union Geriatric Medicine Society. All rights reserved.

Available online at

www.sciencedirect.com

1. Introduction

Cognitive impairment and dementia have been found to beassociated with driving cessation [1,2]. However, a noteworthyminority of cognitively impaired older drivers continue to drivedespite their impairment [3]. Cognitive impairment can have anadverse affect on driving. For example, Alzheimer’s disease andmild cognitive impairment (MCI) has been found to be associatedwith a decline in driving performance [4,5] and increased drivingdifficulty [6] which leads to concerns about traffic safety.

Self-regulation of driving is in general regarded as a strategy forcontinuing to drive safely despite functional decline (e.g. [7]. Olderdrivers with functional decline have been found to self-regulatetheir driving by reducing their overall amount of driving andavoiding challenging driving situations [8,9] and they tend to self-regulate their driving more than well-functioning older drivers[3,8,10–12]. However, older drivers do not always respond tofunctional decline by regulating their driving [13]. It has been

* Corresponding author. Technical University of Denmark, Department of

Transport, Bygningstorvet 115, 2800 Kgs, Lyngby, Denmark.

Tel.: +4535324895, +4545256518; fax: +4545936533.

E-mail addresses: anm@transport.dtu.dk, annette.meng@psy.ku.dk (A. Meng),

aks@transport.dtu.dk (A. Siren), tom.teasdale@psy.ku.dk (T.W. Teasdale).

1878-7649/$ – see front matter � 2013 Elsevier Masson SAS and European Union Ger

http://dx.doi.org/10.1016/j.eurger.2013.01.002

proposed that it is the drivers’ self-monitoring of their own drivingcapacity or ability that determines the extent to which theirdriving is regulated [14,15].On the other hand, it has also beenargued that the self-regulation of driving is an automatic processconducted in order to minimise the cognitive load, and that theperson therefore may not be aware that he or she is compensatingfor functional loss [16].

Meng and Siren [17] explored the role of driving-relateddiscomfort in the self-regulation of driving. Their results suggestthat the experience of cognitive problems is associated withdriving-related discomfort and that driving-related discomfortmay function as an indirect monitoring of driving ability amongolder drivers in general. In addition, the study by Meng and Sirenexplored how those older drivers who recognise cognitiveproblems perceived changes in their driving skills using theperspective of hierarchical models of the driving task [18,19]. Thepattern of change in driving skills at the different hierarchicallevels reported by the participants in their study suggests thatthose older drivers who recognise cognitive problems display goodself-assessment of changes in their driving skills, despite theproblems.

The results from the study by Meng and Siren [17] were basedon a population based random sample of older drivers and theresults therefore do not necessarily apply to those older drivers

iatric Medicine Society. All rights reserved.

A. Meng et al. / European Geriatric Medicine 4 (2013) 154–160 155

who are cognitively impaired. Knowledge of the extent to whichthese findings apply to cognitively impaired older drivers mayprovide useful information on how to address the safety concernsraised by their driving.

The aim of the present study was therefore to examine whetherthe findings of the Meng and Siren study would also apply to asample of cognitively impaired older drivers, in order to gain agreater understanding of the self-monitoring of driving ability, andself-assessment of changes in driving skills among such drivers.

2. Methods

The data were collected by structured face-to-face interviewswith 25 cognitively impaired drivers aged 60–92 (mean 74.4 years,SD 7.5). Twenty (80%) were males and five (20%) females. Theirmean Mini Mental Status Examination (MMSE) [20] score was25.0, (SD 2.9) ranging from 18–28. They had completed an averageof 12 (SD 3) years of schooling.

The participants were a convenience sample fulfilling theminimum criteria for inclusion in the study (i.e. having been for acognitive evaluation at a neurological ward or memory clinicconfirming cognitive impairment and driving at least once amonth). They were recruited through neurological wards andmemory clinics at hospitals in Copenhagen and Aarhus inDenmark. Several hospitals in Denmark were provided with adescription of the study and the approval of the study from theethical committee. The hospitals were asked if it was possible forthem to help recruiting participants for the study. If they agreed tohelp, they were provided with information about the study to handto their patients. The role of the hospital staff was solely to handthis information to their patients when they attended the wards orclinics, and to ask if a researcher could contact them. Whenpatients agreed, their contact information was given to the firstauthor who then contacted the patients and arranged theinterviews.

All interviews were conducted by the first author at theparticipants’ own homes in the time period January toNovember 2011. The interviews started with the participantsigning an informed consent form following which thestructured interview was conducted. The interviewer readout each question and noted down the answers. Because manyof the participants had difficulties with memory and concen-tration, the answer categories were presented to the partici-pants visually. Answers to all questions were not obtained forall 25 participants due to participants not knowing the answeror understanding the question. Exact numbers for each itemare indicated below when they were less than 25. Finally, theMMSE was administered. The interviews took between 40 and60 minutes to complete.

The interviews were based on the questionnaire used in theMeng and Siren [17] study. Demographic information includedage, gender, marital status, driving frequency, and annual mileage.Seventy-two percent reported being married, 8% single, and 20%widowed. Their mean annual mileage was 6,997 km (SD 3,925,n = 22). Forty percent reported driving daily, 56% several times aweek, and 4% at least once a month. As an objective measure ofhealth status, the participants were presented with a list of 20symptoms and illnesses and asked whether they suffered fromthese as confirmed by a physician. This list was derived fromprevious studies with similar settings and participants [21]. Inaddition, the participants were asked to assess their subjectiveoverall health on a four-point Likert-type scale of 1 = excellent,2 = good, 3 = fair, and 4 = poor. They reported an average of 2.8 (SD1.7) symptoms and illnesses, and 92% rated their health asexcellent or good.

To get an indication of self-rated cognitive functioning, theparticipants were asked how often they experienced difficultieswith memory, concentration, orientating themselves, working outhow to approach a task, making themselves understood duringconversations, reacting too slowly, or saying something which theylater regretted. They were asked to indicate how often theyexperienced each of these seven difficulties on a five-point Likert-type scale of 1 = very rarely, 2 = rarely, 3 = neither rarely nor often,4 = often, and 5 = very often. In the analysis, the sum of these wascalculated in order to obtain a global self-rated cognitive score. Thescores thus ranged from 7–35; the higher the score, the poorer theself-rated cognitive functioning. Cronbach’s Alpha coefficient wascalculated for this scale and revealed that it had good internalreliability (Cronbach’s a = 0.72). In addition, the participants wereasked to indicate how often they experienced problems with theirvision on the same five-point Likert-type scale used for cognitivefunctioning.

Their mean self-rated cognitive score was 15.4 (SD 4.1, n = 20).The cognitive problems they were most likely to report were withmemory, which 64% reported to experience problems with often orvery often, followed by concentration and orientation, for both ofwhich 16% reported experiencing problems often or very often.Regarding vision, 21% reported experiencing problems with visionneither seldom nor often while the rest of the participants (79%)reported experiencing problems with vision very seldom.

2.1. Changes in driving skills

The participants were presented with a list of 14 driving skillsand asked to assess whether their skills had declined, wereunchanged, or had improved, in comparison with the level 15 yearsearlier. The list of skills was based on models of the driving task[18,19], which divide driving into hierarchical levels representingvarious aspects of the task and the required skills. The lowest levelinclude skills relevant to the immediate traffic situation such asvehicle manoeuvring while the higher levels include skills such asstrategic decisions, self-monitoring, and social interaction intraffic, and relate to, among other, the safety attitudes of thedriver [19]. It was ensured that both higher level and lowest levelskills were represented in the list (see Table 1 for list of skills).

2.2. Discomfort in driving situations

To assess driving-related discomfort, the participants wereasked how unpleasant it was to drive in 20 driving situations on afour-point Likert-type scale of 1 = not unpleasant at all, 2 = a littleunpleasant, 3 = unpleasant, 4 = very unpleasant. The list was basedon previous literature about situations which older drivers tend toexperience as stressful [22,23], as well as on a selection of commondriving situations (see Table 2 for the list of driving situations). Inselecting the situations, account was taken of the hierarchicalmodels of the driving task, and the aim was to include situationswhere operating the vehicle draws upon different driving skills.The situations, ‘times and places with many cyclists’ and ‘rightturns when there are cyclists’, were added because of theexceptionally large population of cyclists in Denmark. A numberof the participants stated that they did not drive in some of thesituations and therefore they were not able to rate their level ofdiscomfort when driving in these situations. This explains theslightly different numbers of responses to the various situations(see Table 2). In the analysis, we combined the answers ‘‘a littleunpleasant’’, ‘‘unpleasant’’, and ‘‘very unpleasant’’ into one‘‘unpleasant’’ and thereby created a dichotomous variable; ‘‘notunpleasant’’ and ‘‘unpleasant’’. In the results section, we thusreport the proportion of the participants who reported that theindividual driving situations were ‘‘unpleasant’’.

Table 1Proportion of the participants from the present study and from the Meng and Siren [17] study reporting the individual skills as declined, unchanged, and improved compared

to 15 years ago.

Skill Cognitively impaired

drivers

(n: 22–25)

Lower functioning group

[17] (n: 406–422)

Higher functioning group

[17] (n: 402–418)

Odds ratiob

Better

%

No

change

%

Worse

%

Better

%

No

change

%

Worse

%

Better

%

No

change

%

Worse

%

CI/LF CI/HF

Better/no

change

Worse/no

change

Better/no

change

Worse/no

change

Avoid taking unnecessary

risks in traffic

36.0 64.0 0 45.6 54.2 0.2 32.1 67.1 0.7

Keep distance to car in front 28.0 72.0 0 51.2 47.6 1.2 37.2 62.6 0.2 2.8

Stop for the pedestrians at

zebra crossings

20.0 80.0 0 31.0 68.2 0.7 20.7 79.0 0.2

Judge whether you are too

tired to drive

18.2 81.8 0 38.2 58.6 3.2 25.6 74.1 0.2

Be aware of cyclists 16.7 75.0 8.3 41.1 57.0 1.9 33.2 66.3 0.5 3.2 15.4

Comply with the traffic rules 16.0 84.0 0 27.5 71.3 1.2 24.6 74.4 1.0

Be patient with other road

users even when they make

mistakes

16.0 84.0 0 44.7 50.7 4.5 34.7 62.2 3.1 4.6

Adjust the speed to the traffic

to avoid breaking hard

12.0 88.0 0 37.4 61.0 1.7 30.0 69.5 0.5 4.5

Judge when it is safe to

enter a larger road from a

smaller side street

12.0 88.0 0 26.0 71.2 2.9 18.5 80.0 1.5

Manoeuvre the car when it is

slipperya

9.1 77.3 13.6 13.6 66.5 19.9 15.7 74.6 9.7

React to unforeseen events in

the traffic

4.3 82.6 13.0 10.3 74.1 15.5 6.4 83.3 10.3

Foresee dangerous situations

in the traffic

4.0 92.0 4.0 23.9 71.5 4.6 24.3 74.2 1.5 7.7 7.5

Timely place the car

appropriately on the road

according to the route ahead

4.0 88.0 8.0 31.2 66.2 2.6 24.3 74.7 1.0 10.4 7.1

Know the traffic rules 4.0 72.0 24.0 7.1 72.9 20.0 8.5 81.2 10.4

a n: 382 for both the lower and the higher functioning group regarding responses to this skill.b Odds ratios calculated as above unity, where Fisher’s Exact Test was significant (P < 0.05, two-tailed).

A. Meng et al. / European Geriatric Medicine 4 (2013) 154–160156

2.3. Self-regulation of driving

Finally, the participants were asked if they avoided each of 17challenging driving situations whenever possible. This list wasbased on previous literature about situations older drivers tend to

Table 2Proportion of the participants in the present study and from the Meng and Siren [17]

Driving situation Cognitively impaired drivers

(n: 18–25)

Low

[17]

% %

Driving when tired 76.9a 84.4

Driving when feeling unwell 66.7b 82.9

Driving when it is slippery 58.3 88.7

Unknown routes 54.5 67.7

Unknown places 54.5 67.4

When it is dark 54.2 78.7

Long trips 40.9 47.0

Motorway 34.8 35.1

Right turns when there are cyclists 32.0 48.9

Times and places with dense traffic 32.0 69.8

Times and places with many cyclists 28.0 65.9

Junctions without traffic lights 28.0 48.7

Left turns 20.0 38.2

Listening to the radio while driving 16.7 39.2

Driving fast 14.3 42.9

Being overtaken by others 12.5 42.0

Having a conversation while driving 8.3 52.1

Overtaking 8.3 49.4

When not having right of way 4.3 22.5

Roundabouts 4.0 20.3

a n:13.b n: 9.c Odds ratios calculated as above unity, where Fisher’s Exact Test was significant (P

avoid [7,8,10,11,24–30] as well as on a selection of commondriving situations (see Table 3 for list of driving situations). Again,account was taken of the hierarchical models of the driving task inorder to include situations where operating the vehicle draws upondifferent driving skills.

study reporting the individual driving situations as ‘‘unpleasant’’ to drive in.

er functioning group

(n: 152–422)

Higher functioning group

[17] (n: 94–418)

Odds ratioc

% CI/LF CI/HF

59.9

56.4

74.0 5.6

39.5

40.6

47.8 3.1

21.6

19.4

29.3

42.3 4.9

38.1 5.0

27.6

11.0

27.8

21.1 4.5

26.5 5.1

27.0 12.0

25.4 10.8

8.2 6.4

6.0

< 0.05, two-tailed).

Table 3Proportion of the participants in the present study and the Meng and Siren [17] study reporting to avoid the individual driving situations and their mean number of situations

avoided.

Driving situation Cognitively impaired drivers

(n: 20–25)

Lower functioning group

[17] (n: 410–422)

Higher functioning group

[17] (n: 404–418)

Odds ratiob

% % % CI/LF CI/HF

Driving when tired 91.7 78.8 72.8

Driving when feeling unwella 80.0 81.8 79.9

Driving when it is slippery 64.0 68.7 57.6

Times and places with dense traffic 56.0 45.1 27.3 3.4

Long trips 37.5 37.9 28.0

When it is dark 32.0 50.0 30.5

Unknown places 29.2 30.4 22.8

Listening to the radio while driving 29.2 53.0 46.6 2.7

Driving fast 29.2 45.1 34.1

Unknown routes 25.0 30.0 21.7

Motorway 24.0 19.5 16.3

Overtaking 20.8 32.3 25.2

Having a conversation while driving 16.0 38.4 29.7 3.3

Times and places with many cyclists 4.3 28.0 13.7 8.5

Roundabouts 4.2 2.1 1.9

Junctions without traffic lights 4.0 9.0 3.6

Left turns 0 3.8 2.4

Mean number of situations avoided 5.3 6.6 5.2 – –

a n: 396 for the lower functioning group and 383 for the higher functioning group.b Odds ratios calculated as above unity, where Fisher’s Exact Test was significant (P < 0.05, two-tailed).

A. Meng et al. / European Geriatric Medicine 4 (2013) 154–160 157

2.4. Statistical analysis

The results of the present study are compared with selectedresults from the Meng and Siren [17] study, which was based on alarge random sample of the older driver population in Denmark. Inthat study, the participants were categorised into the higher andthe lower functioning participants based on their global self-ratedcognitive score. The median score (10) was used to divide theparticipants into the two groups. The higher functioning group hada mean score of 7.8 and the lower functioning group a mean scoreof 13.0. The participants in the present study (the cognitivelyimpaired drivers) had a mean score of 15.4 and thus on averagereported the most problems with their cognitive functions of thethree groups.

Fisher’s Exact Test was used to test the statistical significance ofthe differences between the groups. Because we wanted to exploredifferences between the cognitively impaired drivers and thelower functioning group and between the cognitively impaireddrivers and the higher functioning group separately, as well asdifferences between these regarding improvements and decline inskills separately, we divided the data into 2 � 2 contingency tablesin the analyses. To test for correlations between constructs, wecalculated the Kendall’s Tau.

3. Results

3.1. Change in driving skills

The majority of the participants reported their skills asunchanged compared to 15 years ago. When reporting change,it was more often improvement than decline of skills. Improve-ment of skills was most often reported for ‘‘avoid takingunnecessary risks in traffic’’ and ‘‘keep distance to the car infront’’. Decline of skills was most often reported for ‘‘know thetraffic rules’’, ‘‘manoeuvre the car when it is slippery’’, and ‘‘react tounforeseen events in the traffic’’. When comparing with theparticipants in the Meng and Siren [17] study, there was atendency for a larger proportion of the cognitively impaired olderdrivers to report that their skills had not changed. In particular thelower functioning group was more likely to report improvement ofskills than the cognitively impaired drivers (Table 1).

3.2. Discomfort in driving situations

The driving situations that the largest proportion of theparticipants reported discomfort in were: driving when feelingtired and when feeling unwell, followed by when it is slippery,unknown routes and unknown places, and when it is dark. Whencomparing with particularly the lower functioning participants inthe Meng and Siren [17] study, there was a tendency for a smallerproportion of the cognitively impaired drivers to report some levelof discomfort in the driving situations (Table 2).

3.3. Avoidance of traffic situations

The participants reported avoiding a mean of 5.3 situations(n = 15, SD 2.4). The driving situations that the largest proportionof the participants reported avoiding were driving when feelingtired and when feeling unwell, followed by when it is slippery, andwhen it is dark. Compared to particularly the lower functioningparticipants in the Meng and Siren [17] study there was a tendencyfor a smaller proportion of the cognitively impaired driversreporting to avoid the driving situations. However, for the drivingsituations when tired, dense traffic, and motorway, a largerproportion of the cognitively impaired drivers than both of theother groups reported avoiding the situation. The differences,however, mostly did not reach the level of statistical significance(Table 3).

3.4. Associations between constructs

We calculated the Kendall’s Tau correlation between discom-fort and avoidance of the driving situations. As shown in Table 4,there is a consistent pattern of positive, mainly significant,correlations between discomfort in a driving situation andavoidance of the driving situation. It should be noted that someratings for discomfort were lacking because participants avoidedthe situation in question entirely. It was not possible to calculatethe Kendall’s Tau correlation for left turns because none of theparticipants reported to avoid left turns.

In addition, we calculated the Kendall’s Tau for discomfort in adriving situation and MMSE score and for discomfort in a drivingsituation and self-rated cognitive score. There was only one

Table 4Kendall’s Tau correlation between discomfort in a driving situations and avoidance

of the driving situation (n: 18–25).

Driving situation t

Motorway/highway 0.39

Junction without traffic lights 0.28

Times and places with many cyclists 0.33

Roundabouts 1.00**

Dense traffic 0.44*

Long trips 0.79**

Unknown routes 0.61**

Unknown places 0.47*

Dark 0.58**

Slippery 0.56**

Driving fast 0.19

Left turns n/a

Overtaking 0.67**

When tired (n: 12) 0.0

When feeling unwell (n: 8) 0.25

Listening to the radio while driving 0.98**

Having a conversation while driving 0.80**

* P < 0.05.** P < 0.01

A. Meng et al. / European Geriatric Medicine 4 (2013) 154–160158

significant correlation (of 17) between discomfort in the drivingsituation and scores on the MMSE, namely –0.45 with ‘‘drivingfast’’. Three correlations with self-rated cognitive scores weresignificant, namely ‘‘junctions without traffic lights’’ 0.44, ‘‘listen-ing to the radio while driving’’ 0.48, and ‘‘having a conversationwhile driving’’ 0.42. One other correlation was significant, namely‘‘driving when it is slippery’’, but this was in the reverse direction,namely –0.47.

Furthermore, we explored the association between the MMSEscore and avoidance of the driving situations and the self-ratedcognitive score and avoidance of the driving situations. There wereno statistically significant correlations between these measures.

Finally, there was no association between the MMSE score andthe self-rated cognitive score (t = 0.02).

4. Discussion

The aim of the present study was to examine whether thefindings of the Meng and Siren [17] study would also apply to asample of cognitively impaired older drivers, in order to gain agreater understanding of the self-monitoring of driving ability, andself-assessment of changes in driving skills among such drivers.

There were four main findings. First, the participants were mostlikely to report their driving skills as unchanged compared tofifteen years ago, and more likely to report no change in their skillsthan, in particular the lower functioning participants in the Mengand Siren study. The driving skills most likely to be reported asimproved were ‘‘avoid taking unnecessary risks in traffic’’ and‘‘keep distance to the car in front’’ and decline of skills was mostoften reported for ‘‘know the traffic rules’’, ‘‘manoeuvre the carwhen it is slippery’’, and ‘‘react to unforeseen events in the traffic’’.Second, generally a smaller proportion of the participants reporteddiscomfort in the driving situations than particularly the lowerfunctioning participants in the Meng and Siren study. Third,generally a smaller proportion of the participants than ofparticularly the lower functioning participants in the Meng andSiren study reported avoiding the driving situations. Fourth, therewas a clear positive association between driving-related discom-fort and avoidance of driving situations. However, there were onlyfew and weak correlations between cognitive status and driving-related discomfort. Moreover, there was no statistically significantassociation between cognitive status and avoidance of drivingsituations and finally, no statistically significant associationbetween MMSE score and self-rated cognitive score.

A larger proportion of the participants in the present study thanthe participants (particularly lower functioning) of the Meng andSiren study, reported their driving skills as unchanged compared tofifteen years ago. Meng and Siren argue that the increase in thehigher level skills reported by the lower functioning participants intheir study may reflect increased caution in traffic and strategicdriving in response to their cognitive problems. The participants inthe present study were much less likely to report improvement oftheir higher level skills while at the same time fewer than could beexpected, reported a decline of the lowest level skills. Thesefindings could be interpreted as indicating that the participants inthe present study lacked insight into their own impairments.Holland and Rabbitt [11] found that many older drivers were notaware of their functional decline, and cognitive impairment is insome cases associated with anosognosia (e.g. [31]. However, theparticipants in the present study reported the most cognitiveproblems suggesting that they do recognise such problems asmemory deficits, but largely do not feel that it affects their driving.Driving cessation is mostly associated with negative consequencesamong older drivers [32,33] and driving is regarded as a necessityand integral part of their independence [26]. The results couldtherefore reflect that they downplay, or are reluctant to acknowl-edge, the impact of cognitive impairments on their drivingperformance because they fear having to give up driving. Analternative explanation may be that they despite recognisingcognitive problems simply lack insight into changes in theirdriving skills. We did, however, not have an objective measure ofthe changes in the participants’ driving skills and therefore do notknow to what extent their driving was actually affected by theircognitive impairments.

When reporting change in their driving-related skills, theparticipants were most likely to report improvement of higherlevels skills and when reporting decline it was in the lowest levelskills. This is the same pattern of change reported by the olderdrivers of the Meng and Siren [17] study and is in line withgerontological knowledge of ageing and cognitive development[34]. However, the skill ‘‘timely place the car appropriately on theroad according to the route ahead’’ is at the strategic level and thusa higher level skill, yet it was among the skills reported to havedeclined. Many of the participants experienced problems withorientation which may challenge this particular driving skill. Itthus appears that cognitive impairments, in some cases, may havean adverse effect on higher level driving skills as well, whereas age-related cognitive problems mainly appear to affect the lowest leveldriving skills. Also, ‘‘Know the traffic rules’’ is a higher level skill,but is reported to have declined in both studies. The decline in thisskill may simply reflect the fact that the older drivers may not feelupdated or may not remember all the traffic rules. A largerproportion of the cognitively impaired older drivers than theparticipants of the Meng and Siren study reported decline in thisskill which may be a consequence of the memory problems theyreported, making them more disposed to experience difficultiesrecalling the formal traffic rules.

The results showed an association between discomfort andavoidance of driving situations indicating that driving-relateddiscomfort plays an important role in the self-regulation of driving,not only among older drivers in general, but also amongcognitively impaired older drivers. However, there were onlyfew correlations between cognitive status both regarding MMSEand self-rated cognitive score and discomfort while in the Mengand Siren study, a significantly larger proportion of the lowerfunctioning group reported driving-related discomfort than thehigher functioning group. The participants in the present studywere a relatively small and homogenous group and therefore thislack of association between cognitive status and discomfort maypartly be explained by lack of variance. Nevertheless, a smaller

A. Meng et al. / European Geriatric Medicine 4 (2013) 154–160 159

proportion of the participants in the present study reporteddriving-related discomfort than particularly the lower functioninggroup of the Meng and Siren study. So, while the recognition ofcognitive problems appears to be associated with driving-relateddiscomfort in older drivers in general, the association may be lessclear among cognitively impaired older drivers. In other words, it ismuch less clear what triggers the feeling of driving-relateddiscomfort among cognitively impaired drivers.

There were no correlations between MMSE score and avoidanceof driving situations. Other studies have likewise failed to findassociations between MMSE score and self-regulation of drivingand driving performance [9,35,36]. More surprising was thefinding that there were no correlations between self-ratedcognitive score and avoidance of the driving situations. In theMeng and Siren study a significantly larger proportion of the lowerfunctioning group reported avoidance of driving situations thanthe higher functioning group. Furthermore, the participants of thepresent study reported avoiding an average number of drivingsituations that was more similar to the higher functioning groupdespite having a self-rated cognitive score that was closer to thelower functioning group of the Meng and Siren study. Hence onceagain, while there appears to be an association between therecognition of cognitive problems and self-regulations of drivingamong older drivers in general, the association is much less clearamong cognitively impaired older drivers.

It could be that the older drivers with cognitive impairmenthave high confidence levels despite their impairments andconsequently experience low levels of driving-related discomfort,which again makes them less inclined to regulate their driving.Given the association found between driving-related discomfortand avoidance of driving situations, the relatively low level ofavoidance may partly be explained by their low level of driving-related discomfort. More intriguing is their relatively low level ofdriving-related discomfort. It may be that lack of driver confidenceleads to driving-related stress or discomfort. A longitudinal studyby Baldock et al. [13] showed that a decline in cognitivefunctioning was not associated with a decline in driver confidence.Given the association between driving-related discomfort and self-regulation of driving, future research into the relationship betweenactual driving ability and driving-related discomfort may shedfurther light on the safety implications of the low levels of driving-related discomfort found among cognitively impaired olderdrivers.

Finally, there was no correlation between MMSE score and self-rated cognitive score. Many factors may influence the degree towhich cognitive impairments affect the daily functioning of aperson. The self-rated cognitive score was an expression of howoften the participants experienced problems with various cogni-tive functions, which may to a larger extent represent their dailyfunctioning and to a lesser extent their objective cognitive status.This may be part of the explanation for finding no correlationbetween these two measures.

It should be noted that there was a procedural differencebetween the two studies. Meng and Siren [17] used phoneinterviews with a random sample of drivers drawn from the Danishdrivers license database while in the present study, we used face-to-face interviews with participants recruited through hospitals.This could lead to greater reluctance to report the negative aspectsof their driving because they may have felt less anonymous andmay have worried that their answers would be reported back to thehospitals. However, great effort was made to avoid this effect bystressing that they were anonymous and that the study wasindependent of the hospital and had nothing to do with drivingevaluation or license renewal.

The participants were all volunteers which may have produceda bias towards cognitively impaired drivers who did not experience

changes to their driving skills, which could be another explanationfor why they largely did not report changes to their skills.

A limitation of the study was the small sample and theoperational definition of cognitive impairment. Most of theparticipants did not yet have a diagnosis and therefore the causeof their impairments was not known. Nevertheless, even thoughmany of the participants did not meet the MMSE cut off score fordementia, as a group they scored below the Danish norm of theirage group [37]. Nonetheless, caution should be taken whengeneralising the results to cognitively impaired older drivers ingeneral and older drivers with dementia in particular. Furtherresearch is needed to establish the robustness of the findings.

Finally, we acknowledge the risk of type I errors due to multiplecomparisons in the analyses. However, we elected not to apply anycorrection for this (e.g. Bonferroni) because of the explorativenature of the study and we therefore were more concerned withavoiding type II errors.

5. Conclusion

Overall the results suggest that cognitively impaired olderdrivers largely constitute a unique group of older drivers. Despitehaving a self-rated cognitive score closer to the lower functioningolder drivers of the Meng and Siren [17] study, they were moresimilar to the higher functioning older drivers regarding their levelof driving-related discomfort and self-regulation of driving.

These findings suggest that the cognitively impaired olderdrivers do recognise cognitive problems but largely do not feel thatit affects their driving. In light of the great importance attached todriving, these results could indicate that the older drivers arereluctant to acknowledge, or that they downplay, the impact ofcognitive impairment on their driving. In addition, the resultsshow that cognitive impairment is associated with a somewhatdifferent pattern of change in driving skills than age-relatedcognitive decline.

Furthermore, the results suggest that driving-related discomfortplays an important role in the self-regulation of driving amongcognitively impaired older drivers as well. However, while driving-related discomfort appear to be associated with cognitive problemsamong older drivers in general, it is much less clear what triggers thisfeeling among cognitively impaired older drivers. Thus, while olderdrivers in general show good self-monitoring skills, cognitivelyimpaired older drivers may show less reliable self-monitoring skillsand a different pattern of change in their driving skills. Thesefindings suggest that when addressing the safety concerns regardingcognitively impaired older drivers, focussed intervention aimed atthe cognitively impaired older drivers, rather than older drivers ingeneral, may prove to be a better solution.

Disclosure of interest

The authors declare that they have no conflicts of interestconcerning this article.

Acknowledgements

The financial support from the Tryg foundation is gratefullyappreciated.

References

[1] Lafont S, Laumon B, Helmer C, Dartigues JF, Fabrigoule C. Driving cessation andself-reported car crashes in older drivers: the impact of cognitive impairmentand dementia in a population-based study. J Geriatr Psychiatry Neurol2008;21(3):171–82.

[2] Foley DJ, Masaki KH, Ross GW, White LR. Driving cessation in older men withincident dementia. J Am Geriatr Soc 2000;48:928–30.

A. Meng et al. / European Geriatric Medicine 4 (2013) 154–160160

[3] Freund B, Szinovacz M. Effects of cognition on driving involvement among theoldest old: variations by gender and alternative transportation opportunities.Gerontologist 2002;42(5):621–33.

[4] Duchek JM, Carr DB, Hunt L, Roe CM, Xiong C, Shah K, et al. Longitudinal drivingperformance in early-stage dementia of the Alzheimer type. J Am Geriatr Soc2003;51:1342–7.

[5] Wadley VG, Okonkwo O, Crowe M, Vance DE, Elgin JM, Ball KK, et al. Mildcognitive impairment and everyday function: an investigation of drivingperformance. J Geriatr Psychiatry Neurol 2009;22(2):87–94.

[6] O’Conner ML, Edwards JD, Wadley VG, Crowe M. Changes in mobility amongolder adults with psychometrically defined mild cognitive impairment. JGerontology Psychol Sci 2010;65B(3):306–16.

[7] Donorfio LKM, Mohyde M, Coughlin JF, D’Ambrosio LA. A qualitative explora-tion of self-regulation behaviours among older drivers. J Aging Social Policy2008;20:323–39.

[8] Ball K, Owsley C, Stalvey B, Roenker DL, Sloane ME, Graves M. Drivingavoidance and functional impairment in older drivers. Accid Anal Prev1998;30(3):313–22.

[9] Cotrell V, Wild K. Longitudinal study of self-imposed driving restrictions anddeficit awareness in patients with Alzheimer disease. Alzheimer Dis AssocDisord 1999;13(3):151–6.

[10] Charlton JL, Oxley J, Fildes B, Oxley P, Newstead S, Koppel S, et al. Character-istics of older drivers who adopt self-regulatory driving behaviours. TranspRes F Traffic Psychol Behav 2006;9:363–73.

[11] Holland CA, Rabbitt PMA. People’s awareness of their age-related sensory andcognitive deficits and the implications for road safety. Applied CognitivePsychol 1992;6:217–31.

[12] Ross LA, Clay OJ, Edwards JD, Ball KK, Wadley VG, Vance DE, et al. Do olderdrivers at-risk for crashes modify their driving over time? J GerontologyPsychol Sci 2009;10:1–8.

[13] Baldock MRJ, Thompson JP, Mathias JL. Self-regulation of driving behaviouramong older drivers: findings from a five-year follow-up. Australian RoadSafety Research, Policing and Education Conference, November 2008, Ade-laide, South Australia.

[14] Anstey KJ, Wood J, Lord S, Walker JG. Cognitive, sensory and physical factorsenabling driving safety in older adults. Clin Psychol Rev 2005;25:45–65.

[15] Horswill MS, Anstey KJ, Hatherly C, Wood JM, Pachana NA. Older drivers’insight into their hazard perception ability. Accid Anal Prev 2011;63(6):2121–7.

[16] De Raedt R, Ponjaert-Kristoffersen I. Can strategic and tactical compensationreduce crash risk in older drivers? Age Ageing 2000;29:517–21.

[17] Meng A, Siren A. Cognitive problems, self-rated changes in driving skills,driving-related discomfort and self-regulation of driving in old drivers. AccidAnal Prev 2012;49:322–9.

[18] Michon JA. A critical view of driver behaviour models: what do we know andwhat should we do? In: Evans L, Schwing RC, editors. Human behaviour andtraffic safety. New York: Plenum Press; 1985. p. 485–520.

[19] Hatakka M, Keskinen E, Gregesen NP, Glad A, Hernetkoski K. From control ofthe vehicle to personal self-control; broadening the perspectives to drivereducation. Transp Res F Traffic Psychol Behav 2002;5(3):201–15.

[20] Folstein MF, Folstein SE, McHugh PR. Mini-mental state’’. A practical methodfor grading the cognitive state of patients for the clinician. J Psychiatr Res1975;12:189–98.

[21] Siren A, Hakamies-Blomqvist L, Lindeman M. Driving cessation and health inolder women. J Appl Gerontol 2004;23(1):58–69.

[22] Hakamies-Blomqvist L, Siren A. Deconstructing a gender difference: drivingcessation and personal driving history of older women. J Safety Res2003;34:383–8.

[23] Hakamies-Blomqvist L, Wahlstrom B. Why do older drivers give up driving?Accid Anal Prev 1998;30:305–12.

[24] Anstey KJ, Smith GA. Associations of biomarkers, cognition and self-reports ofsensory function with self-reported driving behaviour and confidence. Ger-ontology 2003;49:196–202.

[25] Baldock M, Mathias J, McLean A, Berndt A. Self-regulation of driving and itsrelationship to driving ability among older adults. Accid Anal Prev2006;38(5):1038–45.

[26] D’Ambrosio LA, Donorfio LKM, Coughlin JF, Mohyde M, Meyer J. Genderdifferences in self-regulation patterns and attitudes towards driving amongolder adults. J Women Aging 2008;20:265–82.

[27] Molnar LJ, Eby DW. The relationship between self-regulation and driving-related abilities in older drivers: an exploratory study. Traffic Injury Prev2008;9:314–9.

[28] Ruechel S, Mann WC. Self-regulation of driving by older persons. Phys OccupTher Geriatr 2005;23:91–102.

[29] Raitanen T, Tormakangas T, Mollenkopf H, Marcellini F. Why do older driversreduce driving? Findings from three European countries. Transp Res F TrafficPsychol Behav 2003;6:81–95.

[30] Rimmo PA, Hakamies-Blomqvist L. Older drivers’ aberrant driving behaviour,impaired activity, and health as reasons for self-imposed driving limitations.Transp Res F Traffic Psychol Behav 2002;5:47–62.

[31] Starkstein SE, Jorge R, Mizrahi R, Robinson RG. A diagnostic formulation foranosognosia in Alzheimer’s disease. J Neurol Neurosurg Psychiatr2006;77:719–25.

[32] Rudman DL, Friedland J, Chipman M, Sciortino P. Holding on and letting go: theperspectives of pre-seniors and seniors on driving self-regulation in later life.Can J Aging 2006;25(1):65–76.

[33] Taylor BD, Tripodes S. The effects of driving cessation on elderly drivers withdementia and their caregivers. Accid Anal Prev 2001;33:519–28.

[34] Stuart-Hamilton I. The psychology of ageing: an introduction. London: JessicaKingsley Publishers; 2006.

[35] Edwards JD, Ross LA, Ackerman ML, Small BJ, Ball KK, Bradley S, et al.Longitudinal predictors of driving cessation among older adults from theactive clinical trial. J Gerontol Psychol Sci 2008;63B(1):6–12.

[36] Frittelli C, Borghetti D, iudice G, Bonanni E, Maestri M, Tognoni G, et al. Effectsof Alzheimer’s disease and mild cognitive impairment on driving ability: acontrolled clinical study by stimulated driving test. Int J Geriatr Psychiatr2009;24:232–8.

[37] Schultz-Larsen K, Kreiner S, Lomholt RK. Mini-Mental Status Examination:mixed Rasch model item analysis derived two different cognitive dimensionsof the MMSE. J Clin Epidemiol 2007;60:268–79.