Nadina B Lincoln Jenny L Taylor Kristina Vella Walter P Bouman Kathryn A Radford
Nadina B Lincoln Jenny L Taylor
Kristina VellaWalter P Bouman
Kathryn A Radford
BackgroundDriving relies on multiple abilities, (e.g. memory, judgement, psychomotor abilities) which can be affected by dementia. In the UK “If you have had, or currently suffer from a medical condition or disability that may affect your driving you must tell the Driver and Vehicle Licensing Agency (DVLA).” (http://www.dvla.gov.uk)There is currently a lack of consistency in the methods used to determine whether a person with dementia is considered safe or unsafe to drive (Donnelly & Karlinsky 1990; BPS 2001)
N B Lincoln - Spring 2009
Previous workLincoln et al. (2006) developed a cognitive test battery for predicting safety to drive in people with dementia
The predictive equation was 88% accurate
N B Lincoln - Spring 2009
Current Study ‐ AIMS1. To establish whether assessments found to
be predictive of on‐road performance in previous study are predictive in an independent sample
2. To determine whether the predictive validity can be improved by including other measures
N B Lincoln - Spring 2009
ProcedureReferral: letter or phone call
Invitation letter to participant: including consent form and information sheet
Home Visit: discuss the study and administer cognitive assessments (1 or 2 sessions)
Driving Assessment: with specialised driving instructor
Report: to referrer, GP, and participantN B Lincoln - Spring 2009
AssessmentsGeneral level of function: MMSE
Attention: SDSA Dot Cancellation; Stroop
Executive abilities: SDSA Square Matrices & Road Sign Recognition; BADS Rule Shift & Key Search; Trail Making
Visuospatial perception: VOSP Incomplete Letters, Position Discrimination & Number Location
Memory: SORT Words Immediate; SORT Words Delayed
Information processing: AMIPB Info Processing A
N B Lincoln - Spring 2009
N B Lincoln - Spring 2009
N B Lincoln - Spring 2009
N B Lincoln - Spring 2009
1. Starting Precautions 13. Passing static vehicles 2. Moving Off 14. Merging with traffic stream3. Observation at Side Junctions 15. Anticipating pedestrians4. Observations to rear 16. Anticipating beh. of others5. Use of Signals 17. Use of speed6. Positioning on Road 18. Obedience of road signs7. Cornering 19. Observing speed limits8. Braking 20. Courtesy 9. Accelerating 21. Emergency stop10. Overtaking 22. Reversing11. Driving on roundabouts 23. Parking vehicle12. Driving in traffic lanes 24. Turning right
25. Following too close
Road Assessment
N B Lincoln - Spring 2009
Road AssessmentClassified:
Definitely Safe Probably Safe Probably UnsafeDefinitely Unsafe
Inter-rater reliability in small sample (n=6)
100% agreement in classification between our driving
instructor and an independent instructor. N B Lincoln - Spring 2009
Recruitment122 people
referred
75recruited
47 not recruited
stopped driving
n=4
notinterested
n=29
licence at DVLA
n=7
completedfull assessment
n= 65
completedonly cognitiveassessment
n=10
othern=7
N B Lincoln - Spring 2009
ParticipantsAge Mean 75.2, SD 6.7, Range 59‐88
GenderMen 50 (77%)Women 15 (23%)
Type of dementiaAlzheimer’s 51%Vascular 21%Other 27%
N B Lincoln - Spring 2009
Driving Outcome
0
5
10
15
20
25
30
Definitely Safe Probably Safe Probably Unsafe DefinitelyUnsafe
28
24
12
1
Lower proportion of unsafe drivers than in previous studies
N B Lincoln - Spring 2009
The AIMS1. To establish whether assessments found to
be predictive of on‐road performance in previous study are predictive in an independent sample
N B Lincoln - Spring 2009
Was the old equation predictive?71% correctly classified (n=62)
Predicted group membership (cognitive tests)
Unsafe Safe
Road Pass 8 41
Assessment
Fail 3 10
These individuals are those who are actually unsafe, but would have been missed by the equation
Sensitivity for safe drivers=80.4%; Specificity=25.0%; PPV=83.7%; NPV=23.1% N B Lincoln - Spring 2009
The AIMS1. To establish whether assessments found to
be predictive of on‐road performance in previous study are predictive in an independent sample
2. To determine whether the predictive validity can be improved by including other measures
N B Lincoln - Spring 2009
Could we produce a better equation?
From a hierarchical DFA using the original tests, plus VOSP Number Location we developed a new predictive equation
N B Lincoln - Spring 2009
Results with new equation?83.9% correctly classified (n=62)
Predicted group membership (cognitive tests)
Unsafe Safe
Road Pass 8 41
Assessment
Fail 11 2
These individuals are those who are actually unsafe, but would have been missed by the equation
Sensitivity for safe drivers=95.3%; Specificity=57.9%; PPV=83.7%; NPV=84.6%
These individuals are less of a concern – we would simply recommend that they are tested on the road
N B Lincoln - Spring 2009
Who did the equation miss?
1. Often drove an automatic car (whilst in Hungary) but tested in manual which drives in UK
2. Other psychiatric problems which may have hindered driving
Two were actually unsafe, but predicted to be safe
This highlights the importance of clinician involvement
N B Lincoln - Spring 2009
Preliminary conclusionThe original equation was only moderately predictive in an independent sample
HOWEVER the battery of tests was still predictive
Further analysis is required to establish how best to combine these tests into a reliable predictive equation.
N B Lincoln - Spring 2009
Practical ApplicationDiagnosed with dementia/licence for reviewCognitive Screening Assessment
Continue driving
Referred for on-road assessment at
Regional Mobility Centre
Score more than a select cut-off
Score less than a select cut-off
N B Lincoln - Spring 2009
For further [email protected]
Lincoln, N. B., Radford, K. A., Lee, E. & Reay, A. C. (2006) The assessment of fitness to drive in people with dementia. Int J Geriatr Psych, 21, 1044-1051.
N B Lincoln - Spring 2009