Cognitive, On-road, and Simulator-based Driving Assessment after Stroke Megan A. Hird, HBSc,*† Abeiramey Vetivelu, HBSc,*† Gustavo Saposnik, MD,*‡x and Tom A. Schweizer, PhD*k{

Driving is a complex activity that requires intact cognitive, behavioral, and motor function. Stroke is one of the most prevalent neurologic impairments and can affect all of these functions. However, diagnosis of stroke is not a definitive indicator of driving impairment. Determining fitness to drive after stroke is a very complex process and is typically based on cognitive assessments, on-road performance, simulatorbased assessment, or a combination of the three. The aim of this review was to provide (1) a systematic review of the literature on cognitive, on-road, and simulator assessment after stroke, and (2) address the existing limitations and inconsistencies in stroke and driving research. Our results indicated that of 1413 total stroke patients, 748 definitively passed and 367 definitely failed an on-road assessment, with minimal information provided about clinical presentation. In addition, although the Stroke Driver Screening Assessment, the Useful Field of View Test, and the Rey-O Complex Figure test may have some utility in predicting driving performance, most cognitive measures have been inconsistently and minimally explored. Several limitations were observed across studies such as procedural inconsistencies, including outcome variables used (eg, driving cessation and pass/fail classification) and the heterogeneity of patient samples (eg, time since stroke and stroke location). Due, in part, to the larger variability in results of cognitive, on-road, and simulator-based assessment, there is no consensus regarding a valid and reliable driving assessment for physicians. Future studies should assess poststroke driving fitness by differentiating different stages, severities, and locations of stroke. Key Words: Stroke—driving—driving simulation—on-road—cognitive—assessment. Ó 2014 by National Stroke Association

Introduction Stroke is one of the leading causes of death and disability in North America.1-3 More than two thirds of individuals exhibit some degree of cognitive impairment poststroke,4-6 and with the aging of the population, the number of drivers and patients with From the *Neuroscience Research, St. Michael’s Hospital, Toronto, Ontario, Canada; †University of Toronto, Canada; ‡Stroke Research Unit, Mobility Program, St. Michael’s Hospital; xMedicine, St. Michael’s Hospital, University of Toronto; kDepartment of Surgery, Neurosurgery Division, University of Toronto; and {Institute of Biomaterials and Biomedical Engineering, University of Toronto. Received February 19, 2014; revision received May 21, 2014; accepted June 10, 2014. This work was supported by a Grant-in-Aid and a Personnel Award from the Heart and Stroke Foundation of Canada and an Early Researcher Award from the Ontario Ministry of Research and Innovation to Dr. Tom Schweizer and a Distinguished Clinician

2654

stroke and cognitive impairment is expected to increase substantially. Cognitive impairment is one of the major factors that influences driving performance. For many individuals, the ability to drive is an important source of independence and quality of life7; however, approximately 48% of stroke

Scientists Award from the Heart and Stroke Foundation of Canada to Dr. Gustavo Saposnik. G.S. and T.A.S. equally contributed to qualify as senior authors. Address correspondence to Tom A. Schweizer, PhD, Keenan Research Centre of the Li Ka Shing Knowledge Institute, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8, Canada. E-mail: [email protected]. 1052-3057/$ - see front matter Ó 2014 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2014.06.010

Journal of Stroke and Cerebrovascular Diseases, Vol. 23, No. 10 (November-December), 2014: pp 2654-2670

DRIVING ASSESSMENT AFTER STROKE

patients do not receive driving advice, and 87% do not have a formal assessment of driving ability.8 The impairments associated with stroke, such as visual field defects, hemiplegia, impairments in visual-spatial ability, attention, and executive function,1,4,5 may preclude individuals from driving poststroke. In cases where impairments are more subtle and may be compensated for by other cognitive–behavioral functions, determining fitness to drive is much more challenging. Evaluating driving fitness in patients with neurologic deficits, specifically those who have experienced a stroke, as well as addressing their needs and the safety of the community, has been a significant challenge for health professionals.9 Limited information is available in the literature regarding fitness to drive after stroke. There is a need for a driving assessment method with high reliability and validity that is able to accurately determine whether individuals poststroke are fit to drive or whether their licenses should be revoked or restricted. Our objective was to (1) provide a systematic review on the 3 most common methods of driving assessment (ie, cognitive, on-road, and simulator) after stroke and (2) address the current limitations and inconsistencies in stroke and driving research.

Methods Search Strategy Three reviewers (M.A.H., T.A.S., and G.S.) conducted a literature search of MEDLINE, EMBASE, and PsycInfo in July 2013. Computer searches based on keywords (both individuals and/or in combination) were conducted including the following: ‘‘stroke,’’ ‘‘driving,’’ ‘‘assessment,’’ ‘‘performance,’’ ‘‘road,’’ ‘‘neuropsychological,’’ and ‘‘simulator’’. References from previously retrieved articles were also searched. The search yielded 197 possible studies.

Study Selection Peer-reviewed articles that were published within the past 15 years (January 1998-July 2013) and had used cognitive/neuropsychologic measures, on-road tests, and/or simulator technology to assess the driving performance of stroke patients were included. The following articles were excluded: (1) articles not relevant to driving and/or stroke (n 5 113); (2) non-English publications (n 5 6); (3) case/pilot studies (n 5 6); (4) studies that used a method of assessment (eg, self-report and caregiver indices) other than cognitive tests, on-road evaluations, or simulator scenarios (n 5 26); (5) review articles (n 5 10); (6) editorials, commentaries, or replies (n 5 5); (7) studies that compared cognitive assessments with cognitive assessments (n 5 2); and (8) studies that pooled patient populations together (n 5 13). Data were extracted to a form that included the following information: first author, year of publication,

2655

study population characteristics, driving assessment or fitness method, and results.

Results The literature search yielded 22 articles that met inclusion criteria, of which 16 involved cognitive assessment, 17 involved on-road assessment, and 3 involved simulator assessment. In general, there was a high degree of heterogeneity of patients across and within studies in terms of stroke location, time since onset of stroke, stroke type, history of neurologic impairment, presence or absence of visual impairment, years of driving experience, and so forth (Table 1). For example, 82% of studies10-27 (n 5 18) reported the presence or absence of visual defects, such as visual field loss (eg, quadrantanopia and hemianopia) and neglect, in patient samples. Of the studies reporting quadrantanopia and/or hemianopia (n 5 12), 58% excluded all patients presenting with these defects,11,16,19,20,22,26,27 whereas 42% included patients who presented with these impairments and were deemed able to safely perform a driving evaluation13-15,21,25 based on an expert assessment (eg, ophthalmologist). None of the studies that assessed individuals with visual impairment reported the outcome of these individuals on the driving assessment (eg, pass/fail classification, fitness to drive, and so forth).

Cognitive Assessment Multiple studies attempted to determine which cognitive measures are predictive of the driving performance of stroke patients by correlating cognitive scores with on-road performance,10-16,22,23,25,26,28,29 simulator performance,30 or driving status17,24 (Table 2). Twelve studies have reported that cognitive tests are predictive of driving,10-12,14,15,17,22-26,29 whereas 5 studies report little or no predictive value.14,16,22,28,30 The Stroke Driving Screening Assessment (SDSA) was developed as a screening assessment for drivers poststroke and contains 4 tests: (1) Dot Cancellation Test, (2) Square Matrices Directions, (3) Square Matrices Compass, and (4) Road Sign Recognition Test (refer to Lincoln, Radford, & Nouri).31 Results suggest that the SDSA is relatively successful in predicting pass/fail classification of an on-road evaluation (P , .0510; 78.9% agreement with the principal evaluator13; sensitivity, 71.4%1079.3%13; specificity, 77.8%10,13), although further replication is required. Lundberg et al11 and Selander et al28 investigated the ability of the Nordic version of the SDSA to predict on-road performance. Results suggested that the Nordic version of the SDSA is not as accurate as the SDSA in predicting driving performance (Dot Cancellation, P , .0511; Directions, P , .000111; Compass, P ,.000111; Road Sign Recognition, P ,.000111; sensitivity, 48%28; specificity, 76%28).

Author

Patient sample size

Time since onset

Stroke (n 5 66), on-road (n 5 43)

Median 5 56 (10-2190) d

Lundberg et al11

Stroke (n 5 97)

Ponsford et al12

Stroke (n 5 200)

First site (n 5 17): range 5 3-9 mo, second site (n 5 58): 1.1 6 1.45 y (range 5 .1-8.0), third site: N/A Median 5 12 mo

Akinwuntan et al13

Stroke (n 5 38)

6-15 w

Akinwuntan et al14*

Stroke (n 5 68)

Akinwuntan et al15

Stroke (n 5 43)

65.9 6 8.4

Driving experience (y) Mean 5 47.3 (range 5 5-76)

History of neurologic impairment N/A

N/A

Ischemic, hemorrhagic (intracerebral and subarachnoid), TIA (n 5 4)

63.0 6 12.45

N/A

First site (n 5 17): no prior history of stroke, other sites: N/A

Median 5 62.0 (16-85)

N/A

Cardiovascular disease (n 5 116), new stroke before assessment (n 5 1), seizures (n 5 14)

53.9 6 12.8

32.1 6 12.7

Mean 5 15 6 18 mo

53 6 13

33 6 13, 24,000 km/y

Median 5 9 (7-20) mo

55 6 12

33 6 11, 24,000 6 27,000 km/y

Stroke type

N/A

Side of lesion Right (n 5 24), left (n 5 37), other (n 5 5)

Right (n 5 40), left (n 5 49), bilateral (n 5 1), N/A (n 5 7)

N/A

No prior history of stroke

Ischemic (n 5 26), hemorrhagic (n 5 12)

Right (n 5 20), left (n 5 16), bilateral (n 5 2)

No history of epilepsy

Ischemic (n 5 53), hemorrhagic (n 5 15)

Right (n 5 35), left (n 5 31), bilateral (n 5 2)

Number of strokes: one Ischemic (n 5 22), (n 5 29), two (n 5 10), hemorrhagic (n 5 21) three (n 5 4); history of epilepsy (n 5 5)

Right (n 5 19), left (n 5 22), bilateral (n 5 2)

Visual defects Included: patients required to have 120 horizontal vision for on-road test Excluded: patients with homonymous hemianopia

Vision (satisfactory, n 5 160; questionable, n 5 20), impaired visual attention (n 5 37), impaired stereopsis (n 5 25), significant visual field loss (n 5 11), reduced visual field (n 5 16), not fulfilling visual requirements for driving (n 5 19) Included: hemi or quadrantanopia (n 5 10); able to perform evaluation based on ophthalmologist assessment (n 5 5), visual acuity of 20/20 (n 5 22), 20/25 (n 5 12), 20/30 (n 5 4) Included: hemianopia (n 5 10), quadrantanopia (n 5 4) Included: Hemi or quadrantanopia (n 5 8), visual acuity of 20/20 (n 5 25), 20/25 (n 5 12), 20/30 (n 5 6)

M.A. HIRD ET AL.

George and Crotty10

Stroke patient age (y)

2656

Table 1. Summary of stroke patient characteristics

Stroke (n 5 441)

12.3 6 20.9 mo

65.4 (15.4)

N/A

N/A

N/A

Sommer et al19y

Stroke (n 5 109)

Stroke: 12.39 6 20.24, (units not specified)

51.39 6 8.92

Median annual distance 5 20,000 km/y

N/A

Stapleton et al20

Stroke (n 5 46), on-road test (n 5 35) Stroke (n 5 27)

Median 5 2 (1-23) mo

63.5 6 13.4

N/A

No visual impairment, coronary heart disease, neuronal disease (eg, seizures), mental disorders, kidney disease, substance abuse, medication that impairs driving No history of epilepsy

60.0 6 13.6

Range 5 17-61

Mazer et al22

Stroke (n 5 84)

Mean 5 10.4 6 15.8 mo, median 5 4.8 (1.0-96.0) mo

60.8 6 11.9

N/A

No history of class IV cardiac status, uncontrolled seizures

N/A

Right (n 5 38), left (n 5 45), bilateral (n 5 1)

Klavora et al23

Stroke (n 5 56)

Minimum 6 mo

60.2 (44-82)

N/A

No history of heart failure, uncontrolled seizures, uncontrolled diabetes, brain stem injury, psychiatric or substance abuse problem, dementia, physical inability to execute motor sequences

N/A

Right (n 5 34), left (n 5 18), bilateral (n 5 4)

N/A

N/A

Ischemic (n 5 9), hemorrhagic (n 5 18)

N/A

Right (n 5 12), left (n 5 15)

Excluded: patients with driving related issues (eg, macular degeneration) Excluded: visual acuity (with corrective lenses) $.5 diopter in 1 eye and $.2 diopter in the other eye, visual impairment (eg, hemianopia, neglect) Excluded: history of hemianopia Included: quadrantanopia (n 5 2), visual acuity of 20/20 (n 5 16), 20/25 (n 5 9), 20/30 (n 5 2) Excluded: visual homonymous hemianopia, primary visual impairment not corrected by lenses Excluded: poor vision

(Continued)

2657

Chua et al18y

Included: normal visual acuity (n 5 228), returned (n 5 142), stopped (n 5 86)

DRIVING ASSESSMENT AFTER STROKE

Returned: right (46%), left (44%), bilateral (9%); stopped: right (51%), left (43%), bilateral (3%) Right (n 5 171), left (n 5 122), bilateral (n 5 5), N/A (n 5 143) N/A

Stroke (n 5 290), returned to driving (n 5 177), stopped driving (n 5 113)

14 6 8.5 mo

N/A, (Median 5 32,500 km/y) No history of stroke, drug or alcohol abuse, psychiatric or medical condition (eg, epilepsy) Returned: 62 6 14, N/A N/A stopped: 68 6 14

Returned: hemorrhagic (8%), stopped: hemorrhagic (19%)

Perrier et al17

1y

54 6 8.8

Right (n 5 14), Excluded: history of left (n 5 10), hemianopia or inadequate bilateral (n 5 3), vision N/A (n 5 6), brain stem (n 5 1)

Stroke (n 5 34)

Akinwuntan et al21

Median 5 6.2 (1.4-14.0) mo

Ischemic (n 5 20), hemorrhagic (n 5 14)

Soderstrom et al16y

Author

Patient sample size

Time since onset

Stroke patient age (y)

Stroke (n 5 50)

Minimum 6 mo

Akinwuntan et al25

Stroke (n 5 104)

18.5 6 20 mo (80-3407 d)

Korner-Bitensky et al26

Stroke (n 5 269)

6.9 6 11 mo

63.6 6 12.5

Lundqvist et al27

Stroke (n 5 30)

8.6 (3-14) mo

68.3 6 4.8

29 (97%) had license for over 10 y

Selander et al28

Stroke (n 5 76)

$6 mo

65.3 6 9.8

N/A

George et al29

Stroke (n 5 24)

Kotterba et al30y

Stroke (n 5 32), MCA (n 5 24), vertebrobasilar (n 5 8)

McKay et al35

Stroke (n 5 30)

Fisk et al

24

Median 5 83.5 (21-816) d

59 6 1.9

Driving experience (y) 109 6 18 miles/wk

56.8 6 11.9 (30-79) 34.9 6 12.4 (6-63) y

N/A

65.6 6 13.2

46.3 6 13.4

7-14 d

54.9 6 14.8

31.0 6 10.1

46.0 6 65.1 mo, range 5 3-280 mo

54.3 6 9.1

N/A

2658

Table 1. (Continued ) History of neurologic impairment

Stroke type

N/A

N/A

No history of epileptic seizures in the last 6 mo

N/A

No history of Class IV cardiac status or uncontrolled seizures No history of epilepsy, severe aphasia, apraxia, or motor impairment

N/A

Ischemic (n 5 26), hemorrhagic (n 5 4)

N/A

N/A

N/A

N/A

TIA (n 5 12), No history of cerebral ischemic stroke neurologic (n 5 20) disorder, cardiovascular disorder, snoring, daytime sleepiness, or medication No history of psychiatric N/A diagnosis or neurologic disorders

Side of lesion Right (n 5 21), left (n 5 21), bilateral (n 5 6), N/A (n 5 2), hemiparesis: right (n 5 24), left (n 5 21) Right (n 5 59), left (45)

Visual defects Excluded: visual disability

Included: visual field problems (n 5 37), hemianopia (n 5 6) Left (42%), Excluded: hemianopia, right (58%) a primary visual impairment not corrected with lenses Right (n 5 5), Excluded: hemianopia, left (n 5 10), included: reduced bilateral (n 5 5), vision (n 5 5) unknown (n 5 10) Right (n 5 35), N/A. Assessment left (n 5 28), of visual acuity and bilateral (n 5 12), visual fields N/A conducted, (n 5 1) results not reported N/A. VSA used, Right (n 5 13), no report of deficits left (n 5 7), other (n 5 4) MCA: right N/A (n 5 11), left (n 5 13)

Right (30%), left (56.7%), bilateral/other (13.3%)

N/A

M.A. HIRD ET AL.

Abbreviations: MCA, middle cerebral artery; N/A, not available; TIA, transient ischemic attack; VSA, Visual Scanning Analyzer (a measure of visual field scanning/loss). *Some patient information additionally reported in Akinwuntan et al.15 yReported locations: Soderstrom et al16: Frontal (n 5 6), frontoparietal (n 5 4), frontal and parietal (n 5 2), frontal and temporal (n 5 1), temporoparietal (n 5 1), frontotemporoparietal (n 5 1), parietal (n 5 1), occipital (n 5 1), thalamus (n 5 1), basal ganglia (n 5 7), cerebellum (n 5 2), brain stem (n 5 1), N/A (n 5 6). Kotterba et al30: MCA (n 5 24), vertebrobasilar (n 5 8). Chua et al18: not specified (n 5 377), occlusion and stenosis of MCA (n 5 29), subarachnoid hemorrhage (SAH, n 5 16), cerebral infarction (n 5 10), other cerebrovascular disease (n 5 6), intracranial hemorrhage (n 5 2), other nontraumatic intracranial hemorrhage (n 5 1). Sommer et al19: ischemic attack (n 5 3), cerebral hemorrhage (n 5 14), SAH (n 5 9), cerebral infarction (n 5 86).

Author

Patient sample size

George and Crotty10 Stroke (n 5 66), on-road (n 5 43)

Tests used

Outcome measure

UFOV, SDSA

On-road pass/fail

On-road pass/fail

Lundberg et al11

Stroke (n 5 97)

NorSDSA

Ponsford et al12

Stroke (n 5 200)

DA,z executive function, Track test, on-road test memory, orientation, language, praxis, visual perception

Akinwuntan et al13

Stroke (n 5 38)

SDSA

On-road pass/fail

Findings SDSA: P , .05; sensitivity 5 71.4%; specificity 5 77.8% UFOV: DA, P , .05; sensitivity 5 85.7%; specificity 5 69.4%; correctly classified 77.5% Selective attention, P , .01; sensitivity 5 42.9%; specificity 5 88.9% Original function*: sensitivity 5 70%; specificity 5 67%; correctly classified 5 68% Discriminant functiony: sensitivity 5 74%; specificity 5 77%; correctly classified 5 75.5% Pass/fail: Dot cancellation time, misses, and false alarms, P , .05; Directions, P , .0001; Compass, P , .0001; Road Sign Recognition Test, P , .0001 Borderline pass/fail: Compass, P , .05; Directions, P , .01 Track test: visual acuity and track speed (r 5 .24, P , .005) and track positioning (r 5 .19, P , .02); visual inattention and track positioning (r 5 .28, P 5 .001); track performance and all tests, r 5 .26 (orientation, P , .002) 2.45 (praxis, P , .001) On-road test: on-road and all tests, r 5 .18 (verbal memory) –..40 (DA, executive, praxis, cognitive rating) praxis, DA, Decision Time and on-road: correctly classified 5 75%; PPV 5 88% Correctly classified 5 78.9% (principal evaluator judgment) (Continued)

DRIVING ASSESSMENT AFTER STROKE

Table 2. Summary of studies reporting cognitive assessment results

2659

Author

Patient sample size

Tests used

Stroke (n 5 68)

Rey-O, UFOV, TAP, SDSA

Akinwuntan et al15

Stroke (n 5 43)

Rey-O, visual neglect

Soderstrom et al16

Stroke (n 5 34), control (n 5 20)

TMT-B, Reaction Time Test, Finger Tapping Test, WCST, Rey-O, Digit Symbol

Akinwuntan et al

Outcome measure

Findings

Rey-O: NS (on-road); r 5 2.40, P , .001 (group decision) UFOV: NS (on-road and group decision) TAP: DA correct response: r 5 .28, P , .05 (on-road), r 5 2.30, P , .01 (group decision) DA median visual RT: NS (on-road and group decision) scanning mean RT: NS (on-road); r 5 .23, P , .05 (group decision) scanning omissions: NS (on-road and group decision) incompatibility SD RT: r 5 2.26, P , .05 (on-road), NS (group decision) incompatibility difference in error: r 5 2.36, P , .01 (on-road), r 5 .28, P , .05 (group decision) Visual Field lateralized mean RT, difference in RT and difference in omissions: NS (on-road and group decision) Visual Neglect lateralized mean RT: NS (on-road), r 5 .35, P , .01 (group decision) Visual Neglect difference in RT: NS (on-road), r 5 .23, P , .05 (group decision) Visual Neglect difference in omissions NS (on-road), r 5 .25, P , .05 (group decision) SDSA: dot cancellation errors: r 5 2.22, P , .05 (on-road), r 5 .30, P , .05 (group decision) Dot cancellation false positive: NS (group decision and on-road) Directions: r 5 .39, P , .01 (on-road), r 5 2.31, P , .05 (group decision) Compass: r 5 .32, P , .01 (on-road), r 5 2.32 (P , .01) Road Sign Recognition: r 5 .22, P , .05 (on-road), r 5 2.30, P , .05 (group decision) Discriminant functionx: sensitivity 5 79.4%; specificity 5 94.1%; PPV 5 93.1%; accuracy 5 86.8% Predriver assessment-pass/fail Discriminant functionx: correctly classified 5 86%; sensitivity 5 77%; specificity 5 92% On-road pass/fail All tests 5 NS Predriver assessmentgroup decision, on-road

M.A. HIRD ET AL.

14

2660

Table 2. (Continued )

Klavora et al23

Stroke (n 5 56)

Fisk et al24

Stroke (n 5 50), controls (n 5 105)

MMSE

Driving status (return to driving poststroke)

MMSE score was related to driving resumption, odds ratio (95% confidence interval) 5 1.3 (1.1-1.4), 30% increase in odds of driving for each 1-point increase on the MMSE MVPT: PPV 5 86.1%; NPV 5 58.3%; score, P 5 .005; MVPT, TMT-A, TMT-B, On-road pass/fail time, P 5 .01 Single Cancellation, TMT-B (,3 errors): PPV 5 85.2%; NPV 5 48.1%; time, Double Cancellation, P 5 .003; errors, P 5 .0002 Road Map TMT-A (,1 error): PPV 5 80.0%; NPV 5 41.9%; time, P 5 .05; errors, NS Single Cancellation (,5 errors): PPV 5 78.9%; NPV 5 44.6%; errors, P 5 .05; time, P 5 .03 Double Cancellation (,5 errors): PPV 5 64.9%; NPV 5 42.6%; errors and time, NS Road Map (,4 errors): PPV 5 72.1%; NPV 5 52.6%; time, P 5 .007; errors, NS DPAB (SDT, DDT, On-road pass/fail CBDI: prediction accuracy 5 66%; false negatives 5 30% CDT, EDT), CBDI (fail the CBDI, but pass the on-road test); false positives 5 4% SDT: prediction accuracy 5 66%; false negatives 5 30%; false positives 5 4% DDT: prediction accuracy 5 68%; false negatives 5 28%; false positives 5 4% CDT: prediction accuracy 5 68%; false negatives 5 28%; false positives 5 4% EDT: prediction accuracy 5 75%; false negatives 5 18%; false positives 5 7% CBDI 1 EDT: prediction accuracy 5 100% Participants who scored #47 on the CBDI were 5.56 times more likely to pass the on-road test than those who scored .47. Participants who scored $195 on the EDT were 7.90 times more likely to pass the on-road test than those who scored ,195 UFOV, BIT Driving Habits Questionnaire Difference between drivers and nondrivers: UFOV, P , .001; BIT, P , .004 (Continued)

DRIVING ASSESSMENT AFTER STROKE

Mazer et al22

Stroke (n 5 290), returned to driving (n 5 177), stopped driving (n 5 113) Stroke (n 5 84)

Perrier et al17

2661

Author

Patient sample size

Tests used

Outcome measure

Stroke (n 5 104)

Rey-O, UFOV, Fimm–Zimmermann test

Final Group Decision based on predriver assessment (suitable/not immediately suitable/not suitable)

Korner-Bitensky et al26

Stroke (n 5 269)

MVPT

On-road pass/fail

Selander et al28

Stroke (n 5 76)

NorSDSA

On-road pass/fail

George et al29

Stroke (n 5 24)

VRST, VSA, RTM

On-road pass/fail, require lessons

Kotterba et al30

Alertness (computer Stroke (n 5 32), MCA test with flashing (n 5 24), vertebrobasilar light-RT, decision (n 5 8), control (n 5 12) making)

Akinwuntan et al

25

Simulator

2662

Table 2. (Continued ) Findings Best predictors predriver assessment: Rey-O, r 5 .42, P , .001; UFOV, r 5 2.43, P , .001 Fimm–Zimmermann test: DA correct response, r 5 .40, P , .001; mean reaction time in scanning, r 5 2.41, P , .001; absolute difference in reaction time in visual neglect, r 5 2.43, P , .001 Best predictors of on-road: Rey-O, r 5 .48, P , .001; UFOV, r 5 2.38, P , .001 Fimm–Zimmermann test: DA correct response, r 5 .39, P , .001; mean reaction time in scanning, r 5 2.40, P , .001; absolute difference in reaction time in visual neglect, r 5 2.38, P , .001 MVPT scores #24: 28 (72%) failed the on-road test MVPT scores of .30: pass rate . fail rate MVPT 35 or 36: 82% passed on-road test MVPT score #30 and .30: PPV of on-road evaluation, 60.9% (range, 48-76%); NPV, 64.2% (range, 52-82%) MVPT cutoff score of 25 had highest PPV (74% of patients who scored #25 failed the on-road test (n 5 32/43) Sensitivity 5 48%; specificity 5 76%; correctly classified 5 62% road sign recognition: P 5 .044 VRST: P , .05 RTM: pass/fail, P , .05 VSA: lessons/fail, P , .05 NS

M.A. HIRD ET AL.

Abbreviations: BIT, Behavioral Inattention Test; CBDI, Cognitive Behavioral Driver’s Inventory; CDT, complex Dynavision task; DA, Divided Attention; DDT, difficult (device paced) Dynavision task; DPAB, Dynavision Performance Assessment Battery; EDT, endurance Dynavision task; MMSE, Mini Mental State Examination; MVPT, Motor-Free Visual Perception Test; N/A, not available; NorSDSA, Nordic version of the Stroke Drivers Screening Assessment; NPV, negative predictive value; NS, results did not reach statistical significance; PPV, positive predictive value; predriver assessment, (medical examination, visual and neuropsychologic tests, on-road test); Rey-O, Rey Osterrieth; RT, reaction time; Complex Figure; RTM, Response Time Measures; SD, standard deviation; SDSA, Stroke Drivers Screening Assessment; SDT, simple (self-paced) Dynavision task; TAP, test for attentional performance battery; TMT-A and TMT-B, Trail Making Test Part A and Part B; UFOV, Useful Field of View; VRST, Visual Recognition Slide Test; VSA, Visual Scanning Analyzer; WCST, Wisconsin Card Sorting Test; WMS, Wechsler Memory Scale. *Original Discriminant Function: Dot Cancellation time, Dot Cancellation false alarms, Compass, Road Sign Recognition. yDiscriminant Function: Dot Cancellation time, Dot Cancellation misses, Compass, Road Sign Recognition. zDA(Telephone search dual task), Executive functions (verbal fluency, reversed digit span, bilateral movements and sequencing), Memory (WMS Digit span forwards, associate learning task, visual and auditory memory), Visual Perception (fragmented letters, embedded figures tests, road map test), Praxis (copying, miming and object use tests), Receptive Language (token test part five), Orientation (WMS Orientation subtest). xDiscriminant Function: visual neglect, Rey-O and on-road test.

DRIVING ASSESSMENT AFTER STROKE

2663

The Trail Making Test Part A and Part B (TMT-A and TMT-B), the Useful Field of View (UFOV), and the Rey-O Complex Figure test have been used to assess driving fitness in numerous patient populations, including stroke.10,14,15,22,24,25 The TMT is a test of attention and visuomotor tracking.1 Although results of Mazer et al22 support the ability of TMT-A and TMT-B to predict performance (pass/fail) on an on-road evaluation (TMT-A time, P 5 .05; TMT-B time, P 5 .003; TMT-B errors, P 5 .0002), Soderstrom et al16 found no correlation between TMT-B and on-road performance. The UFOV is a computer-based test of visual processing and visual attention, and it contains 3 subtests: processing speed, divided attention, and selective attention. George and Crotty10 reported that the UFOV Divided Attention and Selective Attention subtests were significantly associated with the pass/fail classification of the on-road test (P ,.05 and ,.01, respectively). Furthermore, UFOV Divided Attention had the highest sensitivity (85.7%) and correctly classified 77.5% of stroke patients, and UFOV Selective Attention had the highest specificity (88.9%), suggesting that the UFOV is a relatively valid measure of driving ability poststroke.10 Similar results were obtained by Fisk et al24 and Akinwuntan et al,25 who reported that the UFOV was predictive of driving status (ie, drivers vs. nondrivers, P , .001),24 a 3-class decision (unfit/not suitable, temporarily unfit/not immediately suitable, and fit/suitable; r 5 2.43, P , .001),25 and on-road performance (r 5 2.38, P , .001).25 However, the results of Akinwuntan et al14 suggested that the UFOV was not

associated with an on-road assessment or 3-class decision.14 The Rey-O Complex Figure is a test of visual– spatial constructional ability and visual–spatial memory. Akinwuntan et al14,25 reported a relatively high and significant correlation between 3-class decision and the Rey-O Complex Figure (r 5 2.40, P , .00114; r 5 .42, P , .00125). Although Akinwuntan et al25 found a significant correlation between the Rey-O Complex figure and on-road performance (r 5 .48, P , .001), both Akwinwuntan et al14 and Soderstrom et al16 found that this measure was not associated with on-road performance (r 5 .1914). Although a few cognitive tests have been shown to be associated with driving fitness, there is evidence suggesting variability and inconsistencies in results across studies. Several cognitive assessments assess many of the cognitive domains that have been shown to be critical for certain aspects of safe driving (eg, TMT-B and attention, Rey-O Complex Figure and visual–spatial abilities)32; however, further research is needed to establish and confirm the utility of these measures to assist in the screening and assessment of driving fitness.

On-road Assessment At present, on-road assessments are frequently considered to be the most accurate method of driving assessment.33 Of 1413 total stroke participants, 748 (range, 24%-84%) definitively passed and 367 (range, 0%-76%) definitively failed an on-road assessment (Table 3).10-13,16,18-20,22,26-29 The remaining participants

Table 3. Pass/fail classification results of on-road performance, number (%) Author

Total, N

Pass, n (%)

Fail, n (%)

Other classifications

George and Crotty10* Lundberg et al11 Ponsford et al12 Akinwuntan et al13 Akinwuntan et al14y Akinwuntan et al15y Soderstrom et al16 Chua et al18

43 97 135 38 68 43 34 441

36 (83.72) 41 (42.37) 84 (62.22) 9 (23.68) 34 (50.00) 26 (60.47) 19 (55.88) 197 (44.70)

7 (16.28) 30 (30.93) N/A 29 (76.30) 34 (50.00) 17 (39.53) 15 (44.12) 39 (8.80)

Sommer et al19 Stapleton et al20 Mazer et al22 Akinwuntan et al25y Korner-Bitensky et al26 Lundqvist et al27 Selander et al28 George et al29

109 35 84 104 269 28 76 24

85 (77.98) 27 (77.14) 33 (39.29) 41 (39.42) 145 (53.90) 14 (50.00) 50 (65.79) 8 (33.00)

24 (22.02) 0 (.00) 51 (60.71) 18 (17.31) 124 (46.10) 14 (50.00) 26 (34.21) 8 (33.00)

N/A Remaining: borderline pass (n 5 23); borderline fail (n 5 3) N/A N/A N/A N/A N/A Remaining: downgraded license (lessons & reassessment) (n 5 110); downgraded license (lessons w/modification & reassessment) (n 5 55); conditional pass w/restrictions (n 5 40) N/A Remaining: restricted to local area only (n 5 8) N/A Remaining: not immediately suitable (n 5 45) N/A N/A N/A Remaining: required lessons (n 5 8)

Abbreviation: N/A, not available. *Pass classification included those who passed and those who were recommended lessons. yPass/fail classification based on predriver assessment (on-road, neuropsychologic, and visual assessments combined) excluded from analysis.

Author

Patient sample size

Outcome measure

George and Crotty10 Lundberg et al11

Stroke (n 5 66), on-road (n 5 43) Stroke (n 5 97)*

Classified: pass (pass 1 recommended lessons)/fail, standardized route and scoring43 Classified: pass/borderline pass/borderline fail/fail, official scoring criteria44,45

Ponsford et al12

Stroke (n 5 200), ORT and track test (n 5 135)

Classified: pass/? (not specified) track test: adjust speed, car control, positioning ORT: attention, compliance and anticipation (3-point scale: poor, good, very good)

Akinwuntan et al13

Stroke (n 5 38)

Classified: ‘‘fit to drive’’ (pass), ‘‘temporarily unfit,’’ or ‘‘definitely unfit’’ (fail). (1) A principal (CARA) evaluator, (2) research investigator used an adapted TRIP with 49 subitems on 4-point scale (poor performance 5 1, fair performance 5 2, sufficient performance 5 3, good performance 5 4); max score 5 196), (3) state-registered evaluator used a 4-point scale similar to TRIP

Akinwuntan et al14

Stroke (n 5 68)

Soderstrom et al16

Stroke (n 5 34), control (n 5 20)

Classified: 3-class decision (‘‘fit’’, ‘‘temporarily unfit,’’ ‘‘unfit’’) based on predriver assessment on-road: TRIP 49 subitems on a 4-point scale (maximum score 5 196) Classified: pass/fail ORT subitems: maneuvering, attention, placement of the car on the road, speed adjustment, traffic behavior (5-point scale)

Stroke (n 5 441)

Classified: (1) pass, (2) downgraded license (lessons and reassessment), (3) downgraded license (lessons with modifications and reassessment), (4) conditional pass with restrictions, (5) fail (scoring procedure not specified)

Findings 83.72% passed (n 5 36), 16.28% failed (n 5 7) 42.27% passed (n 5 41), 23.71% borderline passed (n 5 23), 30.93% failed (n 5 30), 3.09% borderline failed (n 5 3) Track test (speed and positioning): satisfactory (n 5 87), unsatisfactory (n 5 21), questionable (n 5 27) ORT: 62.22% passed without failures (n 5 84) track test and ORT correlated (r 5 .72; P , .0000) after track test and ORT: 65 judged as ‘‘not OK’’ (unfit), mainly due to cognitive impairment (62%) and visual impairment (28%), 48 judged as ‘‘interim’’ (additional rehabilitation and sometimes adapted car for training), 87 ‘‘OK to drive’’ Inter-rater reliability of overall ORT, ICC 5 .83 Item-per-item reliability, ICC 5 .63-.87 Principal evaluator and research investigator: inter-rater reliability of items on ORT, .44-.78 principal evaluator: 23.68% passed (n 5 9) and 76.32% failed (n 5 29) state-registered evaluator: 5.26% passed (n 5 2), 94.74% failed (n 5 36) State-registered and principal evaluators: classification agreement 5 81.6%, r 5 .80; sensitivity 5 80.6%; specificity 5 100% On-road and group decision, r 5 2.74, P , .001 Predriver assessment: 50.00% failed (n 5 34), 50.00% passed (n 5 34) 44.12% of patients (n 5 15) and 50.00% of the controls (n 5 10) failed the ORT (P 5 NS) Most common reason for failing 5 attention and speed judgment 44.67% passed (n 5 197), 24.94% received a downgraded license (with lessons and reassessment) (n 5 110), 12.47% received a downgraded license (with less with modifications and reassessment; n 5 55), 9.07% conditionally passed with restriction (n 5 40), 8.85% failed (n 5 39)

M.A. HIRD ET AL.

Chua et al18

2664

Table 4. Summary of studies reporting on-road assessment results

Stroke (n 5 109)

Stapleton et al20

Stroke (n 5 46), on-road test (n 5 35)

Akinwuntan et al21

Stroke (n 5 27)

Classified: N/A adapted TRIP checklist: 55 items on a 4-point scale (maximum score 5 220) assessors (n 5 3): 2 from CARA (1 real-life, 1 video); 1 external (assessed video)

Mazer et al22

Stroke (n 5 84)

Klavora et al23

Stroke (n 5 56)

Akinwuntan et al25

Stroke (n 5 104)

Korner-Bitensky et al26

Stroke (n 5 269)

Classified: pass/fail 43-item assessment used 4 sections: controls, maneuvering, specific driving skills (eg, visual exploration and response to traffic signals), general driving skills (eg, decision making, planning, tolerance) Classified: pass/fail based on: knowledge and use of road rules, problem solving, visual processing speed, and risk perception Classified: suitable to drive/not immediately suitable to drive/not suitable Predriver assessment (medical examination and visual, neuropsychologic, and on-road tests) 10 sections (scored 1-4 on each): motor aspects of driving (steering handling, coordination, control rapidity of movement); cognitive and perceptual skills total score: 0-40 Classified: pass/fail pass defined as passing without any restrictions restricted licensing (eg, drive with another person, or on specific routes, or specific times of day) classified as a fail

Classified: pass/fail criteria: (1) lateral distance and distance from lead vehicle, (2) lane keeping, (3) communication with traffic, (4) speed and overtaking, (5) safety behavior, (6) behavior at intersections, (7) orientation behavior, (8) anticipatory driving behavior, (9) operation of vehicle score: (1) not observable, (2) very well handled, (3) handled, (4) error Classified: pass (unrestricted and restricted)/fail (1) JRHREF: 34 skills on a 5-point scale (maximum score 5 170) (2) adapted TRIP checklist: 53 items (50 on a 4-point score and 3 on a 3-point scale; maximum score 5 209)

Stroke: 77.98% passed (n 5 85), 22.02% failed (n 5 24)

All patients who participated passed (n 5 35), 77.14% unrestricted (n 5 27), 22.86% advised to restricted driving to local area (n 5 8) No completion (n 5 11): refused (n 5 6), unsafe (n 5 1), moved jurisdictions (n 5 2), medically deteriorated (n 5 1), no suitable car (n 5 1) All of the test subitems (n 5 55) demonstrated .80% scoring agreement except for 5 items in the real-life vs. video comparison and 3 items in the video vs. video comparison Intraclass correlation coefficients: (overall performance) real-life vs. video, .62; video vs. video, .80 Passed (n 5 33), failed (n 5 51)

DRIVING ASSESSMENT AFTER STROKE

Sommer et al19

N/A

Based on predriver assessment (group decision): 39.4% suitable to drive, 43.3% not immediately suitable, 17.3% not suitable, road test was the best predictor of group decision, r 5 2.67

46.1% (n 5 124) failed

(Continued)

2665

M.A. HIRD ET AL.

Stroke (n 5 76)

Stroke (n 5 24)

Selander et al28

George et al29

Lundqvist et al

Abbreviations: CARA, the Center for Determination of Fitness to Drive and Car Adaptations; JRHREF, Jewish Rehabilitation Hospital Road Evaluation Form; N/A, not available; NS, results did not reach statistical significance; ORT, on-road test; TIA, transient ischemic attack; TRIP, Test for Investigating Practical Fitness to Drive. *Included patients diagnosed with TIA (n 5 4).

33.33% passed (n 5 8), 33.33% required lessons (n 5 8), 33.33% failed (n 5 8)

65.79% passed (n 5 50), 34.21% failed (n 5 26)

50% classified as ‘‘low’’/fail (n 5 14), 50% classified as pass (n 5 14) Patients vs. controls: speed, maneuvering, lateral position, traffic behavior (P , .01); attention (P , .05)

Classified: ‘‘low’’ driving skill (fail) Swedish National Road Administration: (1) speed (in relation to traffic demands), (2) maneuvering (steering, gearing, braking), (3) lateral position, (4) traffic behavior (planning, following traffic rules, adjusting to other drivers), and (5) attention (to road signs and other drivers) Classified: pass/fail based on frequencies and severity of problems (scoring procedure not specified) Classified: pass/require lessons/fail, standardized and scoring procedure46 Stroke (n 5 28), controls (n 5 30)

Outcome measure Patient sample size

27

Author

Table 4. (Continued )

Findings

2666

(n 5 298) received conditional passes (eg, required lessons and reassessment), passes with restrictions, borderline passes/fails, and so forth. The large variability in results is in part because of inconsistencies in methodology across studies such as different courses, evaluators, requirements that determine driving status (ie, checklists), and inclusion criteria (eg, time since injury, stroke location, and stroke type). Akinwuntan et al13,21 evaluated the reliability and validity of on-road assessment (Table 4). The results of Akinwuntan et al21 supported the reliability of the on-road evaluation (ie, for most subitems of the Test for Investigating Practical Fitness to Drive checklist; see Akinwuntan et al,21 Table 4), particularly when assessment was performed with the same evidence (ie, video compared to video recording, intraclass correlation coefficient (ICC) 5 .80 versus real-life compared to video evaluation, ICC 5 .62). Akinwuntan et al13 found support for the validity of on-road assessment using 3 assessors (agreement of pass/fail classification between principal evaluator and state-registered evaluator, 81.6%; sensitivity, 80.6%; specificity, 100%). The results of these studies suggest that on-road assessment is reliable and valid using the Test for Investigating Practical Fitness to Drive checklist. Furthermore, results suggest that although many individuals poststroke are not fit to drive, some retain the ability to drive.10,13,19,20,29

Driving Simulator Assessment Driving simulators have emerged as a promising tool to assess fitness to drive34 that use virtual reality technology. Driving simulators are most often equipped with a standard wheel, pedals, signaling system, and rear-view mirror that can be used to examine a wide range of driving behaviors. Very few studies have investigated the driving performance of stroke patients using driving simulator technology (Table 5). Kotterba et al30 compared the simulated driving performance of middle cerebral artery (MCA) stroke patients (n 5 24), vertebrobasilar stroke patients (n 5 8), and healthy controls (n 5 12). There was no significant difference in accident rates between stroke patients as a whole and controls; however, after removing transient ischemic attack patients (n 5 10) from the analysis, results suggested that patients with complete strokes had significantly more accidents than control participants (P , .05). Patients with MCA strokes had a significantly higher accident rate compared with controls (2.88 6 3.6 vs. 1.25 6 1.36; P , .05) and showed an increased rate in accidents and concentration faults compared with vertebrobasilar stroke patients (P , .05). McKay et al35 also found that stroke patients (n 5 30) performed significantly worse on a simulator evaluation (P , .001) compared with healthy controls (n 5 30). Furthermore, stroke patients overestimated their actual performance more than controls at pretest and post-test, suggesting impaired self-awareness; however, there was evidence

DRIVING ASSESSMENT AFTER STROKE

2667

Table 5. Summary of studies reporting driving simulator assessment results Author

Patient sample size

Outcome measure

Findings

Lundqvist et al27 Stroke (n 5 23), controls (n 5 28) The Swedish Road and Transport Research (VTI) driving simulator Listening span test while driving Variables: speed, lateral position, complex RT, time to collision (TTC), and distance to collision (DTC)

Kotterba et al30

McKay et al35

Patients vs. controls for complex RT, TTC, and DTC, P 5 NS Listening Span test during driving: controls (7.48 6 3.19 words) performed better than patients (4.86 6 2.71 words), P , .01 Simulator performance did not predict on-road classification (NS) Whole stroke group Stroke (n 5 32),* MCA (n 5 24), CAR simulator (stroke 1 TIA): accidents vertebrobasilar (n 5 8), control Highway driving for 60-minute at rates vs. controls, (n 5 12) 100 km/hour with different P 5 NS weather and obstacles stroke (without TIA): Outcome: accidents (with cars, accident rates vs. controls pedestrians, and so forth) and (P , .05) concentration faultsy (headlights MCA: more accidents vs. at wrong time, ignoring speed controls (2.88 6 3.6 vs. limit, and so forth) 1.25 6 1.3), P , .05 more accidents and concentration faults vs. vertebrobasilar stroke patients, P , .05 driving simulator and reported accidents, not correlated, P 5 NS Stroke patients (38.0 6 12.5) Stroke (n 5 30), control (n 5 30) Doron AMOS-2 interactive performed significantly simulator worse than controls (50 6 45-minute: (1) residential light traffic, 10) on the simulator (2) rural traffic and roadways (lane evaluation, P , .001 changes), (3) challenging situations Stroke patients overestimated (eg, near collisions), (4) brake their performance more than reaction, front-end parking, controls (Y self-awareness) distance estimation Stroke patients might Total score: speed, stop distance, benefit from feedback— lane placement, traffic accuracy [ from prediction signal use, hazard (52.4 6 7.5) to postdiction avoidance, and obeying (47.8 6 6.4) signs and signals

Abbreviations: AMOS-2, Advanced Mobile Operation System-2; CAR, Computer Aided Risk; MCA, middle cerebral artery; NS, results did not reach statistical significance; RT, reaction time; TIA, transient ischemic attack. *Included patients diagnosed with TIAs (n 5 12). yConcentration faults were scored manually by a technician.

that suggested that stroke drivers might benefit from feedback (ie, their estimation accuracy increased from prediction to postdiction).35 In contrast to the results obtained by Kotterba et al30 and McKay et al,35 Lundqvist et al27 found no significant difference between patients and controls across several driving simulator variables (eg, time to collision, distance to collision, and complex reaction time). Simulator performance was not associated with on-road performance.27

Discussion There is currently no consensus regarding a valid and reliable assessment method for physicians to accurately determine the driving fitness of their patients. This is, at least in part, because of the large variability in the results of cognitive, on-road, and simulator-based assessments. The variability in the means of evaluation (cognitive, on-road, and simulator-based) across studies explains

M.A. HIRD ET AL.

2668

the observed discrepancy in the results of driving assessment. Results suggest that cognitive tests, specifically the SDSA,10,13 TMT,22 UFOV,10,24,25 and Rey-O Complex Figure14,15,25 may have some utility in predicting driving performance; however, studies are highly variable in terms of which cognitive tests are used to predict driving fitness and numerous measures have been inconsistently and minimally explored. As such, though tests have shown some success, too few studies have supported their reliability and validity in predicting driving performance. It is difficult to translate results into clinical practice and to make recommendations, as no single method has been thoroughly explored and validated. Other factors that may contribute to the aforementioned disparity is the lack of consistent, appropriate, and evidence-based cutoff scores36 as well as the large variability in sample sizes, procedures, and outcome/performance variables across studies. Specifically, outcome variables vary from driving cessation, to pass/fail classification, to 3-class decision, to overall performance in terms of on-road evaluation, simulator performance, predriver evaluations, and selfreport measures. Consequently, until further research is conducted, cognitive measures can only be used to assist decision making regarding driving fitness or when considering referral to on-road assessment.37 The stroke population is highly variable in terms of driving fitness, with some studies reporting a high percentage of participants passing (.75%10,19,20) and others reporting a very low percentage of participants passing (,35%13,29). This variation may, in part, be because of methodological limitations, such as procedural inconsistencies (eg, assessors, administration22 and scoring,22,25 traffic,16 and test routes) across studies and the heterogeneity of patients across and within studies. Although on-road assessments are able to assess a driver’s response to some complex driving situations (eg, highway driving and left turns), responses to many other complex and potentially dangerous driving situations cannot be evaluated (eg, collision avoidance). Drivers with mild impairments may be able to pass more simple aspects of driving (eg, simple turning), but may exhibit deficits when exposed to more complex environments/scenarios (eg, night driving, bad weather conditions, collision avoidance, and high traffic). Such impairments may be undetected in on-road testing. Driving simulator technology permits the examination of driving performance in more risky situations as well as patients’ responses and learning. Driving simulation has been shown to be ecologically valid,38,39 reliable40 across participants, and associated with real-world driving.41 However, very few studies have investigated the driving performance of individuals poststroke. Furthermore, despite the fact that simulator evaluation is consistent across participants within a study, a potential limitation is that there is no standardized hardware, software, sce-

narios, or scoring across studies. Results suggest that stroke patients perform significantly worse on simulated driving scenarios compared with controls30,35 and that they may have impaired self-awareness regarding their deficits,35 whereas transient ischemic attack patients may perform comparably well to controls.30 Results also suggest that MCA stroke patients may have greater impairment in driving poststroke compared with vertebrobasilar stroke patients.30 Although the results of 1 study27 suggested that simulator performance was not associated with on-road performance, the 2 assessments targeted separate aspects of driving. Whereas the onroad assessment evaluated simpler driving behaviors (eg, maneuvering, position, following traffic rules, and attention to road signs), simulator performance evaluated more complex driving behaviors (eg, reaction time, time to collision, distance to collision, and divided attention task). The lack of association between on-road and simulator assessment likely reflects the fact that the evaluations investigated different aspects of driving. Therefore, in general, results support the utility of driving simulator technology poststroke; however, the minimal amount of research precludes the ability to definitely characterize driving behavior poststroke. To address the current limitations in driving assessment research, future studies should combine simulator and neuroimaging technology to investigate the behavioral and neural correlates of driving poststroke as cognitive, motor, and behavioral presentation is highly variable within this population. This would allow researchers to identify which specific cognitive areas are associated with various stages, severities, and locations of stroke influencing driving performance. Ultimately, simulator and on-road procedures, equipment, and evaluations could be standardized based on these results, increasing accuracy and eliminating inconsistencies across studies. Furthermore, isolating which brain areas are associated with specific driving behaviors could lead to the development of more reliable and valid cognitive screening measures. Physicians would be able to perform quick and accurate assessments of patients and refer borderline drivers to driving rehabilitation specialists for more comprehensive onroad42 and simulator-based evaluations. The development of accurate and evidence-based assessments would help restore the autonomy of individuals who are fit to drive and isolate drivers who are unfit to drive, thereby increasing the safety of the driver and the general public. Acknowledgment: The authors thank Kristin Vesely, MSc candidate, for her assistance in the construction of the tables in the article.

References 1. Lezak M. Neuopsychological Assessment. 4th ed. New York: Oxford University Press 2004.

DRIVING ASSESSMENT AFTER STROKE 2. Stasitics Canada. Mortality, Summary List of Causes 2008. 2011. Available at: http://www.heartandstroke. com/site/c.ikIQLcMWJtE/b.3483991/k.34A8/Statistics. htm#stroke. 3. Johnson M, Cohn J, Bakas T. Emergency department nurses’ perceived barriers and facilitators to caring for stroke patients. J Neurosci Nurs 2011;43:238-245. 4. Le
sniak M, Bak T, Czepiel W, et al. Frequency and prognostic value of cognitive disorders in stroke patients. Dement Geriatr Cogn Disord 2008;26:356-363. 5. Nys GMS, van Zandvoort MJE, de Kort PLM, et al. Cognitive disorders in acute stroke: prevalence and clinical determinants. Cerebrovasc Dis 2007;23:408-416. 6. Tatemichi TK, Desmond DW, Stern Y, et al. Cognitive impairment after stroke: frequency, patterns, and relationship to functional abilities. J Neurol Neurosurg Psychiatry 1994;57:202-207. 7. Freund B. Office-based evaluation of the older driver. J Am Geriatr Soc 2006;54:1943-1944. 8. Fisk GD, Owsley C, Pulley LV. Driving after stroke: driving exposure, advice, and evaluations. Arch Phys Med Rehabil 1997;78:1338-1345. 9. Eby DW, Molnar LJ. Driving fitness and cognitive impairment: issues for physicians. JAMA 2010;303: 1642-1643. 10. George S, Crotty M. Establishing criterion validity of the Useful Field of View assessment and Stroke Drivers’ Screening Assessment: comparison to the result of onroad assessment. Am J Occup Ther 2010;64:114-122. 11. Lundberg C, Caneman G, Samuelsson SM, et al. The assessment of fitness to drive after a stroke: the Nordic Stroke Driver Screening Assessment. Scand J Psychol 2003;44:23-30. 12. Ponsford AS, Viitanen M, Lundberg C, et al. Assessment of driving after stroke–a pluridisciplinary task. Accid Anal Prev 2008;40:452-460. 13. Akinwuntan AE, De Weerdt W, Feys H, et al. The validity of a road test after stroke. Arch Phys Med Rehabil 2005; 86:421-426. 14. Akinwuntan AE, Feys H, De Weerdt W, et al. Prediction of driving after stroke: a prospective study. Neurorehabil Neural Repair 2006;20:417-423. 15. Akinwuntan AE, Devos H, Feys H, et al. Confirmation of the accuracy of a short battery to predict fitness-to-drive of stroke survivors without severe deficits. J Rehabil Med 2007;39:698-702. 16. S€ oderstr€ om ST, Pettersson RP, Leppert J. Prediction of driving ability after stroke and the effect of behind-the-wheel training. Scand J Psychol 2006; 47:419-429. 17. Perrier MJ, Korner-Bitensky N, Mayo NE. Patient factors associated with return to driving poststroke: findings from a multicenter cohort study. Arch Phys Med Rehabil 2010;91:868-873. 18. Chua M, McCluskey A, Smead JM. Retrospective analysis of factors that affect driving assessment outcomes after stroke. Aust Occup Ther J 2012;59:121-130. 19. Sommer M, Heidinger C, Arendasy M, et al. Cognitive and personality determinants of post-injury driving fitness. Arch Clin Neuropsychol 2010;25:99-117. 20. Stapleton T, Connolly D, O’Neill D. Exploring the relationship between self-awareness of driving efficacy and that of a proxy when determining fitness to drive after stroke. Aust Occup Ther J 2012;59:63-70. 21. Akinwuntan AE, DeWeerdt W, Feys H, et al. Reliability of a road test after stroke. Arch Phys Med Rehabil 2003; 84:1792-1796.

2669 22. Mazer BL, Korner-Bitensky NA, Sofer S. Predicting ability to drive after stroke. Arch Phys Med Rehabil 1998; 79:743-750. 23. Klavora P, Heslegrave RJ, Young M. Driving skills in elderly persons with stroke: comparison of two new assessment options. Arch Phys Med Rehabil 2000; 81:701-705. 24. Fisk GD, Owsley C, Mennemeier M. Vision, attention, and self-reported driving behaviors in communitydwelling stroke survivors. Arch Phys Med Rehabil 2002;83:469-477. 25. Akinwuntan AE, Feys H, DeWeerdt W, et al. Determinants of driving after stroke. Arch Phys Med Rehabil 2002;83:334-341. 26. Korner-Bitensky NA, Mazer BL, Sofer S, et al. Visual testing for readiness to drive after stroke: a multicenter study. Am J Phys Med Rehabil 2000;79:253-259. 27. Lundqvist A, Gerdle B, Ronnberg J. Neuropsychological aspects of driving after stroke–in the simulator and on the road. Appl Cognitive Psych 2000;14:135-150. 28. Selander H, Johansson K, Lundberg C, et al. The Nordic Stroke Driver Screening Assessment as predictor for the outcome of an on-road test. Scand J Occup Ther 2010; 17:10-17. 29. George S, Clark M, Crotty M. Validation of the Visual Recognition Slide Test with stroke: a component of the New South Wales occupational therapy off-road driver rehabilitation program. Aust Occup Ther J 2008; 55:172-179. 30. Kotterba S, Widdig W, Brylak S, et al. Driving after cerebral ischemia–a driving simulator investigation. Wien Med Wochenschr 2005;155:348-353. 31. Lincoln NB, Radford KA, Nouri FM. Stroke Drivers’ Screening Assessment: revised manual 2012. 2012: 1-25. 32. Schweizer TA, Kan K, Hung Y, et al. Brain activity during driving with distraction: an immersive fMRI study. Front Hum Neurosci 2013;7:53. 33. Driving and Parkinson’s disease. Lancet 1990;336:781. 34. Freund B, Gravenstein S, Ferris R. Evaluating driving performance of cognitively impaired and healthy older adults: a pilot study comparing on-road testing and driving simulator. J Am Geriatr Soc 2002;50: 1309-1310. 35. McKay C, Rapport LJ, Bryer RC, et al. Self-evaluation of driving simulator performance after stroke. Top Stroke Rehabil 2011;18:549-561. 36. Molnar FJ, Patel A, Marshall SC, et al. Clinical utility of office-based cognitive predictors of fitness to drive in persons with dementia: a systematic review. J Am Geriatr Soc 2006;54:1809-1824. 37. Lincoln NB, Taylor JL, Vella K, et al. A prospective study of cognitive tests to predict performance on a standardised road test in people with dementia. Int J Geriatr Psychiatry 2010;25:489-496. 38. Lew HL, Poole JH, Lee EH, et al. Predictive validity of driving-simulator assessments following traumatic brain injury: a preliminary study. Brain Inj 2005;19: 177-188. 39. Lengenfelder J, Schultheis MT, Al-Shihabi T, et al. Divided attention and driving. J Head Trauma Rehabil 2002;17:26-37. 40. Frittelli C, Borghetti D, Iudice G, et al. Effects of Alzheimer’s disease and mild cognitive impairment on driving ability: a controlled clinical study by simulated driving test. Int J Gen Psychiatry 2009;24: 232-238.

2670 41. Lee HC, Cameron D, Lee AH. Assessing the driving performance of older adult drivers: on-road versus simulated driving. Accid Anal Prev 2003;35:797-803. 42. Rapoport MJ, Herrmann N, Molnar FJ, et al. Sharing the responsibility for assessing the risk of the driver with dementia. Can Med Assoc J 2007;177:599-601. 43. George S, Clark M, Crotty M. Development of the Adelaide Driving Self-Efficacy Scale. Clin Rehabil 2007; 21:56-61.

M.A. HIRD ET AL. 44. Swedish National Road Administration. Code of statutes 1996;168. 45. Swedish National Road Administration. Code of statutes 1998;53. 46. Lister R. Driving assessment route and scoring key. Revised by Berndt A & George S. In: Clark MS, Hecker J, Cleland E, et al, eds. Dementia and driving. Canberra, ACT: Australia Transport Safety Bureau 1998:56-64.