Table of Contents
Next Section


II.B. ATTENTION/PERCEPTION/COGNITION


1. Angular Motion Sensitivity

(a) MultiCAD

2. Attentional Search & Sequencing

(a) AARP Reaction Time Test
(b) Auto-Trails
(c) Trail Making Test
(d) WayPoint

3. Attention Switching

(a) Digit Symbol Subscale of WAIS
(b) Washington University Attention Switching Task

4. Attentional Visual Field

(a) Smith-Kettlewell Modified Synemen Perimeter
(b) Visual Attention Analyzer (UFOV)

5. Divided Attention

(a) MultiCAD

6. Driving Knowledge

(a) Rules of the Road
(b) Traffic Sign Recognition

7. Immediate/Delayed Recall

(a) Logical Memory Subscale of Wechsler Memory Scale

8. Language Abilities/Naming Behavior

(a) Boston Naming Test

9. Mental Status

(a) Mattis Organic Mental Status Syndrome Examination (MOMSSE)
(b) Mini-Mental State Evaluation (MMSE)
(c) Short Blessed Cognitive Screen

10. Perceptual Speed

(a) Cue Recognition (Doron Driver Analyzer)

11. Selective Attention

(a) Auditory Selective Attention Test

12. Sustained Attention

(a) Continuous Performance Task

13. Visual Perception

(a) Benton Visual Retention Test
(b) Motor-Free Visual Perception Test
(c) WAIS-R Picture Completion

14. Multiple Capabilities

(a) Cognitive Behavioral Driver's Inventory (CBDI)
(b) Cognitive Screen (DrivAble Testing, Ltd.)
(c) Driving Advisement System (DAS)
(d) Easy Driver
(e) Elemental Driving Simulator (EDS)
(f) University of Iowa/Atari Interactive Driving Simulator
(g) University of Nevada, Las Vegas (UNLV) Subtests
(h) Washington University Visual Attention Tests


ATTENTION/

PERCEPTION/

COGNITION

Angular Motion Sensitivity:

MultiCAD

82 Areferred@ subjects aged 60-91 (26 of which were identified as probably being cognitively impaired to some degree). The drivers were referred to the DMV for reexamination due to a medical condition (by physician, optometrist, ophthalmologist), a series of licensing test failures, a flagrant driving error (police referral), or some other indicator of driving impairment.

This test used MultiCAD to measure drivers' ability to rapidly detect changes in the relative motion of their own versus other vehicles. A video of suburban driving scenes was used which presented a driver's eye view of travel along an arterial route with light traffic, following a lead vehicle (that the subject was told to pay attention to) at varying distances. Subjects were required to depress the "brake" assembly whenever the vehicle directly ahead in the same lane applied its brakes or at any other time it would be advisable to stop or slow down under actual driving conditions (e.g., an adjacent-lane driver encroaches into the lane of travel). The lead vehicle brake lights were illuminated when it slowed for 12 of the angular motion sensitivity trials. For 3 other angular motion sensitivity trials, the lead vehicle=s brake lights were disabled during filming of the video, so that the subject was required to detect the change in headway without the additional brake light cue. These 3 trials were intermingled with the trials in which the brake lights were illuminated.

Measures of effectiveness were: (1) mean brake reaction time across 12 trials, to slowing/stopping lead vehicle with brake light activation, for correct responses; (2) percent error for these trials (e.g. percent of the trials where the vehicle ahead slowed and the brake lights were clearly visible, but the subject did not press the brake pedal; (3) mean brake reaction time across 3 trials, to slowing/stopping lead vehicle with no brake light activation, for correct responses; and (4) percent error for these three trials.

A gross measure was also employed, which was a count of the number of times the word Aerror@ appeared on the printout of results. This measurement ignored any varying stimulus characteristics.

Multiple linear regressions were conducted to arrive at the best linear combination of variables for predicting performance on road tests; (see On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium), and comparisons were made between cognitively impaired and cognitively non-impaired referral drivers to determine whether there were differences in performance on nondriving tests and driving tests.

California DMV Field Office

$A gross measure of the number of errors made in the driving video significantly correlated with weighted error score on the road test (r=.3462, p< .002).

$Mean brake time (in response to a lead vehicle braking with or without brake lights activated) was not significantly correlated with weighted error score on the drive test.

$The correlation between proportion of errors on trials where brake lights were visible and weighted error score on the drive test was significant (r=.2801, p<.013).

$The correlation between proportion of errors on trials where brake lights were not visible and weighted error score on road test was not significant.

$Using gross errors on the driving video, cognitively impaired referral subjects made significantly more errors (average = 7.50) than did the cognitively unimpaired referrals (average = 3.36).

$Looking at the proportion of errors for trials where the brake lights activated, cognitively impaired referrals had a significantly higher error proportion (they did not brake in 47.3% of the trials) compared to cognitively unimpaired referral subjects (who did not brake in 21% of the trials)

$Response time did not discriminate between cognitively impaired referrals and cognitively unimpaired referrals, neither did proportion of errors when the brake lights did not activate (although there were only 3 trials for this last measure).

Janke & Eberhard (1998)

Staplin, Gish, Decina, Lococo, and McKnight (in press)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

AARP Reaction Time Test

3,238 drivers ages 65+, who applied for renewal of North Carolina driver=s license

This test is based on the reaction time test presented in the AARP Older Driver Skill Assessment and Resource Guide, which is similar to Trails A. The test consists of a photo of a driving scene onto which 14 numbers are overlaid; the subject must touch the numbers in order. The test was modified by increasing its size to 15 by 23 in (to widen field of visual search) and instead of timing for 10 seconds and scoring the last number touched, scoring was based on total time to locate and touch all 14 numbers.

Dependent variable: involvement in a police-reported motor vehicle crash during the three-year period immediately preceding license renewal

 

Eight NC driver=s license offices, representing a mix of urban and rural locations in the western, central, and eastern portions of the State.

Performance declined significantly as a function of increasing age (time to complete test increased with increasing age).

Correlational coefficient with number of crashes = 0.046 (p<0.001). Annual crash involvements increased with increasing (poorer) cognitive scores.

Stutts, Stewart, and Martell (1996)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Auto-Trails

$ 69 Areferred@ subjects aged 60-91. The drivers were referred to the DMV for reexamination due to a medical condition (by physician, optometrist, ophthalmologist), a series of licensing test failures, a flagrant driving error (police referral), or some other indicator of driving impairment.

$ 31 paid Avolunteers@ aged 56-85, recruited through signs posted at study site or word of mouth.

A modified and automated version of Trails A of Reitan=s (1958) Trail Making Test, developed by Frank Schieber (Univ of SD). 14 numbers are presented on a computer monitor arranged randomly against the background of a traffic scene, as observed by the driver through the windshield of a car. The subject must touch the numbers (touch screen display) in numerical order as rapidly and accurately as possible. Timing is done by the computer. The score used was total time, as very few subjects made errors.

Three tiers of analyses were conducted in this research: (1) logistic regressions to determine what combination of tests, observations, or survey variables, with what weightings, would best predict whether a subject was a volunteer or referral; (2) multiple linear regressions were conducted to arrive at the best linear combination of variables for predicting performance on road tests; and (3) comparisons were made between cognitively impaired and cognitively non-impaired referral drivers to determine whether there were differences in performance on nondriving tests and driving tests.

(see On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium).

California DMV Field Office

Referral group performed significantly worse than the volunteer group (Correlation between Auto-Trails time and Group = .405, p<.05).

Auto-Trails mean time for referrals = 24.26 s, for volunteers = 16.91 s

Note: this variable was also significantly correlated with age (correlation = .364)

Auto-Trails time correlated significantly with weighted error score on the road test, for combined referrals and volunteers (r=.4523, p<.000) and for referrals only ( r=.3748, p<.002).

Auto-Trails time did not discriminate the cognitively impaired referral subjects from the cognitively unimpaired referral subjects.

A model using number of observed problems + Pelli-Robson errors + Auto Trails time, with a cut-point of p=.86 of being a referral, gave specificity of 96.8% (30 of 31 volunteers classified correctly) with sensitivity of 63.1% (41 of 65 referrals correctly classified). However, the number of subjects in the model was only 96.

A multiple linear regression model using knowledge test score, Auto Trails time, Doron Cue Recognition 2 score, MultiCAD Static Contrast Sensitivity time with the high contrast 20/80 target, and MultiCAD Static Acuity time for correct responses at 20/80 accounted for 56.4% of the variance in performance on the road test (weighted road test error score).

Janke & Eberhard (1998)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Trail Making Test

(Parts A and B)

3,238 drivers ages 65+, who applied for renewal of North Carolina driver=s license

Paper-and-pencil test of general cognitive function. It measures speed of visual search, attention, mental flexibility, and motor function (Reitan, 1958). Part A involves connecting in order 25 encircled numbers randomly arranged on a page. Part B includes both numbers (1-13) and letters (A-L), and requires connecting the two in alternating order (1 to A to 2 to B, etc.). The score on either test is the overall time (seconds) to complete the connections. Mistakes are pointed out by the test administrator and are corrected as they occur; their effect is to increase the overall time required.

Dependent variable: involvement in a police-reported motor vehicle crash during the three-year period immediately preceding license renewal

________________________________________________

FINDINGS (Cont=d)

Of the measures examined, Trails A & B generally performed the best; both are sensitive to milder levels of cognitive impairment. The increase in crash risk observed from the lowest to the highest levels of test performance was very gradual, so that there was no clear cutpoint for identifying a particularly high risk subgroup of drivers.

A multivariate model providing the best fit to the crash data was one containing Trails B time as the cognitive predictor, and self-reported driving frequency, annual mileage, and age as additional explanatory variables. The model accounts for only 3.3% of the total deviance. Estimated 3-year crash totals for selected levels of these 4 variables predict higher crash totals for older drivers (compared to younger drivers) who drive daily as opposed to less often, and for those who drive more miles/yr. Also, subjects who take slightly more than 2 minutes to complete Trails B are at nearly twice the crash risk level.

A second model substituting Trails A for Trails B was almost as strong, as was a third model using Traffic Sign time, driving frequency and age.

Eight NC driver=s license offices, representing a mix of urban and rural locations in the western, central, and eastern portions of the State.

Performance declined significantly as a function of increasing age for both Trails A and Trails B (time to complete test increased with increasing age).

Average completion times were below (better than) published norms, suggesting a healthy or well educated sample.

Trails A Results: Correlational coefficient with number of crashes = 0.065 (p<0.001).

Subjects who scored in best quartile had 47% fewer crashes (.037 crash involvements per year) than drivers who scored in the worst quartile (.054 crash involvements per year).

Trails B Results: Drivers in the poorest decile of performance have a predicted average annual crash rate of 1.5 times that of drivers in the highest decile of cognitive performance. Correlational coefficient with number of crashes = 0.072 (p<0.001). Annual crash involvements increased with increasing (poorer) cognitive scores.

 

Stutts, Stewart and Martell (1996, 1997)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Trail Making Test

(Parts A and B)

105 drivers licensed in Nebraska, aged 65-88 (mean age = 71.4). 54 were females (mean age = 70.5 years); 51 were males (mean age = 72.2 years). All subjects were volunteers, and were paid $25.00 for participating. 36 had taken a driver education course in the past 10 years.

Paper-and-pencil test of general cognitive function. Part A involves connecting in order 25 encircled numbers randomly arranged on a page. Part B includes both numbers and letters, and requires connecting the two in alternating order (1 to A to 2 to B, etc.). The score on either test is the overall time (seconds) to complete the connections. Mistakes are pointed out by the test administrator and are corrected as they occur; their effect is to increase the overall time required.

The driving performance of the subjects was evaluated using the on-street driving performance measurement (DPM) technique developed by Vanosdall and Rudisill (1979). The subjects were evaluated by a driver education expert trained in the use of the DPM technique, while they drove in their own cars. The DPM route was a 19-km circuit designed to evaluate the subjects in the situations that are most often involved in the accidents of older drivers. Therefore, their performance was evaluated at 7 intersections where they were required to make left turns at 5 intersections and right turns at the other 2 intersections. Four of the left turns were made from left-turn lanes onto four-lane divided arterial streets in suburban areas, and one was made from a left turn lane onto a two-lane one-way street in an outlying business district.

Cognitive measures: University laboratory.

Driving measures:

business district and residential street networks

Only the Trail Making Part B test showed a significant correlation to performance on the driving task, with a correlation coefficient of -0.42 (p<.0001). The correlation between Trails A and driving performance was

-0.03 (p<.7329).

The TMB showed the highest correlation of all factors (visual, visual perception, cognitive, range of motion) included in the analysis.

All factors investigated were included in a stepwise procedure of regression analysis. The only significant factors were Trails B, trunk rotation to the right, Trails A, overall visual perception response-time score, and spatial relationship error score, which together accounted for 45 percent of the total variability in driving performance. According to the signs of the regression coefficients in this model, better driving performance was associated with better cognition as measured by the Trail Making Tests, better range of motion in trunk rotation, and better visual perception.

 

Tarawneh, McCoy, Bishu, and Ballard (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Trail Making Test

(Part A)

$ Healthy elderly controls (n=13); mean age = 73.5; CDR score =0

$ Subjects with very mild dementia (n=12) ; mean age = 72.5; CDR score = 0.5

$ Subjects with mild dementia (n=13); mean age = 73.4; CDR score = 1.0

Subjects came from the Washington University Longitudinal Studies population

Dementia severity measured w/ Washington University=s Clinical Dementia Rating

Attentional and visuospatial assessments were conducted prior to the road test. The standard Trails A test was among these pre-driving assessments.

The in-vehicle, on-road driving ability of participants was scored independently by a driving instructor (blind to study design and dementia status of the subjects), and an unblinded occupational therapist. The vehicle was a standard model car w/ automatic transmission and equipped with dual brake pedals. Each subject drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt Apass/fail@ rating was given by each observer in the vehicle.

Washington University Alzheimer=s Disease Research Center.

Five subjects--all in the CDR 1 stage--@failed@ the in-car on-road test. There was 100% agreement between the driving instructor and principal investigator in their pass/fail ratings for all 38 drivers. The ability to follow the driving instructor=s directions, the demonstration of appropriate decision-making (>judgment=) in traffic, and interpretation of traffic signs were highly correlated with overall driving performance. Other behaviors demonstrated by subjects who Afailed@ the in-car exam included coasting to a near stop in the midst of traffic, drifting into other lanes of traffic, stopping abruptly without cause, simultaneously pressing the brake and accelerator while driving, delay in changing lanes when an obstacle appeared, and failure to understand why other drivers signaled them in frustration or exaggeration.

The correlation between the pass/fail outcome on the road test and performance on Trails A was significant at the p<.02 level.

Hunt, Morris, Edwards, and Wilson (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Trail Making Test

(Part B)

121 licensed drivers forming groups composed of :

$ 47 normal/nondemented elderly (mean age 72.9)

$ 29 middle-aged/nondemented controls (mean age 40.6)

$ 45 cognitively impaired drivers (mean age 73.3)

$ 28 with mild dementia

$ 8 with moderate dementia

$ 9 with cognitive impaired but not meeting the criteria for dementia

Paper-and-pencil test of general cognitive function. Part B includes both numbers and letters, and requires connecting the two in alternating order (1 to A to 2 to B, etc.). The score on either test is the overall time (seconds) to complete the connections. Mistakes are pointed out by the test administrator and are corrected as they occur; their effect is to increase the overall time required.

[6 other psychometric tests were included in this study: letter cancellation, stroop, choice reaction time, WAIS-R picture completion, WAIS-R comprehension subtest, and Direct Assessment of Functional Status]

Two operational level dependent measures were collected using the Computerized Driving Assessment Module (CDAM): simulator brake reaction time and simulator steering accuracy. The CDAM consists of an automobile seat, dashboard with speedometer, brake and gas pedals, steering wheel, computer monitor for display of instructions, and a double arc of light-emitting diodes (LEDs) set at eye-level subtending 190 of visual field which generate stimuli for steering tasks.

The brake RT measure comprised the average of three trials, where the subject was instructed to maintain a "speed" of 50 kph while monitoring a screen for the appearance of a STOP sign. RT corresponded to the interval between the appearance of the word STOP and the time the brake pedal was fully depressed. Steering accuracy was computed by summing the areas of deviation between the curve describing the position of computer generated lights and the curve generated by the steering actions of the driver.

Maneuvering level measures were assessed on the Motor Vehicle Branch (MVB) Road Test and on a measure of stopping distance in response to a moving hazard.

Strategical level measures were related to the accuracy of subjects' self appraisals and comprised the Cone Avoidance Task and a comparison between self-ratings and collateral ratings of driving problems. The cone avoidance task required a subject to maneuver a test vehicle through a course of traffic cones, hitting as few as possible.

Cognitive battery given at Clinic for Alzheimer=s Disease and Related Disorders (University Hospital, Vancouver B.C),

CDAM testing performed at a local Rehab Center, MVB Road test conducted by license examiners on a class 5 course.

Cone Avoidance test conducted on off-road course.

Performance on the Trails Test was only significantly correlated with steering deviation performance in the simulator (correlation = .47 p<.05). On the steering deviation task, the demented performed significantly less well than did either of the two control groups. Additionally, the mean steering deviation score for the normal elderly was significantly larger than that of the mid-age group. The only other psychometric measure significantly related to driving performance was WAIS-R picture completion, which was only correlated with brake time on the driving simulator. In both cases, the psychometric tests accounted for less than 25% of the variance in driving behavior.

A note of interest: Although the demented had on average, 10 more demerit points than the normal elderly on the MVB road test, 75% of the demented drivers passed the road test.

There was no significant correlation between these two tests and performance on the motor vehicle branch test, or on stopping distance or cone avoidance.

Tallman, Tuokko, and Beattie (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Trail Making Test

(Part B)

20 drivers age 55+

S=s were primarily recruited through local news media; some were referred by physicians to a driver evaluation service

S=s were administered the following measures:

$ Snellen visual acuity test

$ A continuous performance task (CPT) that used a single letter as a target

$ Simple reaction time

$ 15 items from Boston Naming Test

$ Trail Making Part B

$ Wechsler Memory Test (Mental control and Orientation subtests)

$ A recognition memory test

$ A word fluency test (F-A-S)

$ Embedded Figures Test

$ Visual Search Test

$ Motor Free Visual Perception Test

Dependent measures included the DMV official summary of driver records for each subject and a standard driving assessment that included paper and pencil tests and a behind-the-wheel examination.

Laboratory (Univ. Of Rochester School of Medicine and Dentristy)

Six subjects were classified as below minimum standards in driving performance (a total of 19 or more errors on the NY State Driving Exam). These 6 subjects scored more poorly on Trails B (mean Trails B total time =130.5 s) than the subjects whose on-road driving performance was at least adequate (mean total Trails B time = 93.07 s)

The large interindividual variance resulted in lack of statistical significance for the small sample.

The Trail Making Test (Part B) was the only test that was significantly correlated with driving performance for all subjects (r=0.61, p<0.01).

1 of the 6 below-standard S=s had been involved in a road traffic accident in the past 3 years, and another had a conviction for failure to yield while turning left.

Cushman (1988)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

Trail Making Test

(Part B)

17 subjects (age 57-97; mean age = 75)

6 females and 11 males.

8 S=s were referred from local mental disorder clinics or from local physicians because of possible dementia and associated driving problems.

9 S=s were community residents who did not have suspected dementia or driving problems.

Paper and pencil test where subjects must connect an alternating series of numbers and letters as quickly as possible. The scores are the total time to complete the task and number of errors.

An on-road driving assessment was performed with the subject driving with a certified driving examiner in a dual-brake vehicle. Simple maneuvers were first performed in a parking lot, then subjects joined the flow of traffic and traveled over a prescribed route in moderate to heavy traffic. Subjects were scored on the basis of errors or omissions that correspond to points on the State of New York road test exam; higher scores indicate poorer performance. Therefore a total score was used as well as a determination of whether the subject met or exceeded state standards ("pass") or failed to meet standards ("fail"). In addition, a pass/fail rating was given for the subjects' performance in steering control, braking, acceleration, judgment in traffic, observation skills, and turning skills (particularly left turning).

Clinical tests: University Laboratory

On-road driving evaluation: parking lot and in-traffic (moderate to heavy traffic situations)

Results of the driving exam indicated that eight subjects passed, eight failed (scored 19 or more errors on the on-road exam), and one could not complete the exam because of poor vision. The analyses conducted in this study compared the subjects who met the driving exam standards with the eight who did not. There was no significant difference in average age of subjects who passed the exam compared to those who failed. Drivers who failed drove significantly fewer miles, however. The below standard group took significantly longer to complete the Trail Making Part B test (mean = 266 s for S=s who failed road test; mean =117 s for S=s who passed road test, t(4.5)=3.21, p=0.027).

A regression analysis to determine which variables predict driving status was not possible, because some subjects did not complete all measures and because the sample size was relatively small. An exploratory analysis using total score on the road test as the criterion measure and using five preselected variables determined that age, total time on Trail Making Test, and the number of omission errors on the continuous performance AX test were possible predictors, and when average reaction time is added, account for 93% of the variance in the road test scores.

Of the 8 persons referred for possible dementia, 5 failed the road test, 2 passed the test, and 1 was unable to complete the evaluation.

Cushman (1992)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

WayPoint

(a derivative Trails B procedure)

Six validation studies with 102 drivers age 20-60:

$ passenger vehicle drivers

$ fuel truck drivers

$ transit bus operators

$ long distance tractor trailer drivers

$ emergency response trainees

$ race car drivers

Four-minute, paper-and-pencil test, where subjects connect alternating numbers and letters in sequence.

WayPoint presents 6 exercises in pamphlet form. The first 4 exercises contain 8 numbers and 7 letters which are to be connected in alternating number-letter order by means of a continuous pencil line; the last two exercises contain 5 numbers and 4 letters to be connected in the same way. Some exercises have small pictures used as irrelevant distractors. Subjects are instructed to keep going if they make a mistake. Performance on each exercise is timed with a stopwatch.

Can be administered one-on-one or in a group. Uses a (proprietary) windows-based scoring program to assess accident risk (high or low), and a narrative about the person=s strengths and weaknesses.

 

(1) Emergency Response Course and Non-Emergency Response Vehicle Operations Course at the Federal Law Enforcement Training Center (FLETC), Brunswick, GA.

(2) Road Atlanta Race Course

(3)MARTA -Metropolitan Atlanta Rapid Transit Authority

Based on 207 driver subjects (bus, car, truck, race car drivers), WayPoint correctly classified 72% as high or low accident drivers, missed 18% of the high accident drivers, and falsely labeled 9.2% of the drivers as high accident when they were actually low accident.

Results were interpreted to show that errors on WayPoint were directly related to (1) technical errors on a closed course (a high speed drive circuit), (2) directly related to line-of-travel errors, and (3) positively correlated with lap speed. In addition, technical errors were correlated with WayPoint Focus, a measure of a person=s ABig Picture.@

On the non-emergency test, WayPoint errors were positively correlated with driving errors and with the number of traffic cones contacted on the obstacle course.

Crash frequency was obtained through self-report in some validation studies and through driver record files in other validation studies.

Michael Cantor

WayPoint Research, Inc.

Atlanta, GA

 

ATTENTION/

PERCEPTION/

COGNITION

Attentional Search & Sequencing:

WayPoint

101 licensed drivers (39 females and 62 males) age 72-90 (mean age = 78.3) who were members of a preexisting study cohort engaged in longitudinal studies of a community-dwelling cohort of older people (at Buck Center for Research in Aging)

WayPoint presents 6 exercises in pamphlet form. The first 4 exercises contain 8 numbers and 7 letters which are to be connected in alternating number-letter order by means of a continuous pencil line; the last two exercises contain 5 numbers and 4 letters to be connected in the same way. Some exercises have small pictures used as irrelevant distractors. Subjects are instructed to keep going if they make a mistake. Performance on each exercise is timed with a stopwatch.

In this study, WayPoint was administered twice (in succession) to see if drivers with presumed cognitive impairment either failed to improve from the first administration to the second, or did not improve as much as subjects without presumed cognitive impairment.

The scoring system determines (1) channel capacity or information-processing rate, defined as the average speed per exercise on the first administration over two of the exercises and (2) high vs low risk of preventable and non preventable collisions, reflecting the driver=s situational awareness.

An on-road driving exam was given by the project driving instructor (owner/operator of a driving school in San Francisco) based on the California Driving Performance Evaluation (DPE), and using the same scoresheet as used for the MDPE given in San Jose by these researchers. (see On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium). A weighted error score was calculated as total # of unweighted errors, plus twice the sum of critical and hazardous errors. Concentration errors were also noted.

Critical errors = errors which would in normal circumstances cause test termination (turning from improper lane, dangerous maneuver, examiner intervention needed).

Hazardous errors = dangerous maneuver or examiner intervention.

Concentration errors = subject unable to proceed to field office at end of test, or drove past the street on which the field office was located and did not recognize their error.

 

Novato, Marin County California; Buck Center for Research in Aging

$Average time per exercise on the first administration of WayPoint was strongly related to road test weighted errors (r=.37) as was channel capacity (r=.35). These correlations were significant.

$A substantial but not significant correlation was found between overall average time per exercise on the second administration of WayPoint and weighted road test errors (r=.31).

$The difference in number of errors from WayPoint 1 to WayPoint 2 was not significant, but the difference in time (5.2 s less on WayPoint 2) was significant.

$Two of the 3 cognitively impaired subjects failed to improve their time scores from the first administration to the second.

$A multiple regression model using 98 subjects using age, average time per exercise on WayPoint 1, Perceptual Response Time (Part 1 of the UFOV), and average number of cognitive domains on the MMSE in which subjects made 1 error yielded a significant prediction of weighted error score on the drive test (Multiple R = .484, adjusted R2 =0.202.

$Substituting channel capacity for WayPoint average time reduced the number of subjects to 92, and yielded a multiple R of .475 and adjusted R2 of 0.190.

$Eliminating age and using Waypoint average time, MMSE error areas and PRT yielded a multiple R of .462, and adjusted R2 of 0.188. Substituting channel capacity for average WayPoint time yielded Multiple R of .451, adjusted R2 of 0.176.

Using only WayPoint 1 average time and PRT as predictors of weighted error score on the road test yielded multiple R = .428; adjusted R2=0.166.

Janke and Hersch (1997)

ATTENTION/

PERCEPTION/

COGNITION

Attention Switching:

Digit Symbol Subscale of Wechsler Adult Intelligence Scale

$ Healthy elderly controls (n=13); mean age = 73.5; CDR score =0

$ Subjects with very mild dementia (n=12) ; mean age = 72.5; CDR score = 0.5

$ Subjects with mild dementia (n=13); mean age = 73.4; CDR score = 1.0

Subjects came from the Washington University Longitudinal Studies population

Dementia severity measured w/ Washington University=s Clinical Dementia Rating

The Digit Symbol subscale of Wechsler Adult Intelligence Scale was given to participants prior to the on-road drive test. This symbol substitution task demands rapid switching of attention between different sources of information. In the WAIS booklet, four rows of blank squares are presented, with each square having above it a randomly assigned number form 1 to 9. At the top of the page is a Akey row@ that pairs each number, in order, with a different abstract symbol. Following practice trials, the subject must fill in each blank square with the symbol corresponding to its number. The subject is instructed to do this as quickly as possible. After 90 seconds, the test is terminated, and the subjects score is the number of squares filled in correctly. The Digit Symbol test taps visuomotor coordination, fine motor speed, speed of mental operation, visual short-term memory, and visual incidental learning.

The in-vehicle, on-road driving ability of participants was scored independently by a driving instructor (blind to study design and dementia status of the subjects), and an unblinded occupational therapist. The vehicle was a standard model car w/ automatic transmission and equipped with dual brake pedals. Each subject drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt Apass/fail@ rating was given by each observer in the vehicle.

Washington University Alzheimer=s Disease Research Center.

Five subjects--all in the CDR 1 stage--@failed@ the in-car on-road test. There was 100% agreement between the driving instructor and principal investigator in their pass/fail ratings for all 38 drivers. The ability to follow the driving instructor=s directions, the demonstration of appropriate decision-making (>judgment=) in traffic, and interpretation of traffic signs were highly correlated with overall driving performance. Other behaviors demonstrated by subjects who Afailed@ the in-car exam included coasting to a near stop in the midst of traffic, drifting into other lanes of traffic, stopping abruptly without cause, simultaneously pressing the brake and accelerator while driving, delay in changing lanes when an obstacle appeared, and failure to understand why other drivers signaled them in frustration or exaggeration.

The correlation between the pass/fail outcome on the road test and performance on the Digit Symbol Test was significant at the p<.007 level.

Hunt, Morris, Edwards, and Wilson (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attention Switching:

Washington University Attention Switching Task

$ Healthy elderly controls (n=13); mean age = 73.5; CDR score =0

$ Subjects with very mild dementia (n=12) ; mean age = 72.5; CDR score = 0.5

$ Subjects with mild dementia (n=13); mean age = 73.4; CDR score = 1.0

Subjects came from the Washington University Longitudinal Studies population

Dementia severity measured w/ Washington University=s Clinical Dementia Rating

Attentional and visuospatial assessments were conducted prior to the road test. In attention switching, subjects were given a sheet of paper with rows of randomly intermixed numbers and letters in large print. Subjects were asked to circle only numbers until instructed to switch and circle only letters until again instructed to switch. Alternating commands (letters or numbers) were given every 30 seconds for 2 minutes. Subjects failed this test if they independently switched (i.e., without the command), circled both numbers and letters without discrimination, or were unable to complete the task.

The in-vehicle, on-road driving ability of participants was scored independently by a driving instructor (blind to study design and dementia status of the subjects), and an unblinded occupational therapist. The vehicle was a standard model car w/ automatic transmission and equipped with dual brake pedals. Each subject drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt Apass/fail@ rating was given by each observer in the vehicle.

Washington University Alzheimer=s Disease Research Center.

Five subjects--all in the CDR 1 stage--@failed@ the in-car on-road test. There was 100% agreement between the driving instructor and principal investigator in their pass/fail ratings for all 38 drivers. The ability to follow the driving instructor=s directions, the demonstration of appropriate decision-making (>judgment=) in traffic, and interpretation of traffic signs were highly correlated with overall driving performance. Other behaviors demonstrated by subjects who Afailed@ the in-car exam included coasting to a near stop in the midst of traffic, drifting into other lanes of traffic, stopping abruptly without cause, simultaneously pressing the brake and accelerator while driving, delay in changing lanes when an obstacle appeared, and failure to understand why other drivers signaled them in frustration or exaggeration.

All 5 CDR 1 subjects who failed the road test performed poorly on the attention switching task. The correlation between the pass/fail outcome on the road test and performance on the attention switching task was significant at the p<.0001 level.

Hunt, Morris, Edwards, and Wilson (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Smith-Kettlewell Modified Synemen Perimeter

(Optifield II)

visual field-integrity loss and attentional visual field-integrity loss

3,669 randomly-selected Class C license renewal applicants, licensed in California for at least 12 years, and unable to renew by mail. Four driver age groups were studied:

26-39, 40-51, 52-69, and 70+.

The perimeter looks like half a large globe 2.5 ft in diameter. The subject is seated looking into the globe, to view a small spot of red light at the far end of the globe. For the Standard Visual Field test, subject focuses eyes on red spot and releases a button each time a green light is flashed. Test spots (green lights) are presented 5 times at 8 different distances from the red focal light along each of 5 meridia (e.g., spokes of a wheel). The meridia stretched to the upper right (rear-view mirror location), the far left and the far right, and the lower left and lower right (where lane boundaries would be seen in one=s side vision). For the Attentional Visual Field test, the red fixation light blinked on and off, irregularly. In addition to pressing a button each time a green light appeared, the subject was required to count and remember how many times the red fixation-light blinked.

Both tests must be given in a dimly-lighted room, and each requires about 6 minutes to administer.

5 experimental vision tests were employed:

$ Pelli-Robson Low-Contrast Acuity Test (measures loss in low contrast acuity; ability to see objects and borders)

$ Smith-Kettlewell Low-Luminance Card (measures high-contrast near-acuity loss and low-contrast near-acuity loss)

$ Berkeley Glare Tester (measures low-contrast near acuity loss, and low-contrast near-acuity loss in the presence of glare)

$ Modified Synemen Perimeter (measures standard visual field-integrity loss and attentional visual field-integrity loss)

$ Visual Attention Analyzer (measures loss in UFOV, the area of the visual field in which useful information can be rapidly extracted from a complex visual display)

The dependent measure was the crash frequency during the previous 3-year period, extracted from the DMV database.

Drivers also completed a Driving Habits Survey measuring level of restriction (never, sometimes, often or always) for night driving, rain or fog, sunrise or sunset, driving alone, left turns, and heavy traffic.

California DMV Field Offices:

Carmichael

El Cerrito

Roseville

Standard field: S=s rated test as face valid (clear instructions, safety-related, and fair in requiring driver license applicants to pass similar sensory tests to get full driving privileges).

Attention field: clarity of instructions rated high, but safety-relatedness and fairness of requirement to pass were not rated as high as for sensory tests. Regression analyses showed that S=s who performed more poorly on attentional tests tended to rate them more negatively.

$For all age groups combined, test score was not significantly associated with total prior 3-year crash involvement when considered in isolation.

$Standard field integrity was excellent for all age groups, but, when an attentional task was added, the S=s age 70+ showed a marked deterioration. They also showed high variability in attentional field-integrity loss.

$Poor perf. on the standard field test showed small, but statistically significant predictive value for S=s aged 26-39 and 70+.

$After adjusting for gender, age, and exposure, the standard field test explained 1.9% of the variance for all age groups, and the attention field test explained 1.6% of the variance. Approx. 5% of the variation in reported level of self-restriction was explained by test performance or age (the worse the visual perf. or the older the driver, the more restriction). There was no significant association between vision score and avoiding heavy traffic or driving alone for standard field, or for heavy traffic for attentional field. 11.4% of the variation for age 70+ S=s was accounted for by the avoidance of left turns and low attention-field scores.

Hennessy (1995)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

Drivers are referred by the State DMV (PennDOT) for a driving evaluation

As of 12/8/95, 48 subjects ages 64-84 were screened. Results presented here are for these 48; however as of 10/11/96, 55 subjects were screened, and their results strengthen the trend.

In Progress Study: AEnhancing Mobility in the Older Driver through Improving the Useful Field of View.@

Older drivers are referred (physician or PennDOT) to Bryn Mawr for a driving evaluation. UFOV screening was added to the existing driving protocol, which includes an on-road driving assessment. The drivers who fail the UFOV screening are eligible for recruitment into the UFOV screening study. UFOV will be assessed post-training, and participants will be given a repeat driving evaluation.

A model 2000 Visual Attention Analyzer was used to measure the detection, localization and identification of suprathreshold targets in complex displays. The size of a person=s UFOV is determined by manipulating three variables: target presentation duration, the competing attentional demands of the central and peripheral task, and the salience of the peripheral target. Three subtests provide a measure of the percentage reduction of a maximum 35 degree radius field. During the first subtest (measuring processing speed capability and vigilance), the test participant must identify a centrally located object which varies in duration, by pressing an icon of a truck or a car (whichever was presented) on the touch-screen display. The second subtest (measuring divided attention capabilities) requires the same identification, in addition to locating a simultaneously presented peripheral target of varying eccentricity. A third subtest (measuring selective attention capabilities) required the same two responses required for subtests 1 and 2 while the peripheral target is embedded in distractors. The composite measure of UFOV reduction is recorded as a percentage ranging from 0% to 90%, and the basis for the loss can be determined by considering the percentages of loss on the three subtests.

Bryn Mawr Rehabilitation Center (Pennsylvania)

Results of the UFOV screening are strongly correlated with on-road driving evaluation; of the clients who pass the UFOV test (less than 40% reduction in UFOV), the majority pass the on-road evaluation, and of the clients who fail the UFOV test (have > 40% reduction in UFOV), the majority fail the on-road evaluation.

Of the 23 drivers who passed the UFOV, 18 passed the on-road, 4 failed on-road, and 1 is pending.

Of the 25 drivers who failed UFOV screening, 6 passed the on-road eval., 16 failed the on-road, and 3 are pending.

[It was noted that for many of the older subjects, the UFOV protocol was tedious and tended to undermine their confidence; they carried this loss of confidence over to the driving exam. Therefore the test protocol was changed to that the driving examination was conducted first, and was followed by the UFOV test.]

Results for 9 drivers who received training (8-to-12-sessions) with the UFOV protocol showed improved UFOV scores, but only 2 drivers improved on-road performance.

On-going NIA project sponsored through Roybal Center for Mobility Enhancement in the Elderly.

Principal Investigator - Tom Kalina

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

294 older drivers, ages 56-90 years at enrollment, drawn from the population of licensed drivers in Jefferson County over age 55.

33% had 0 crashes on record

49% had 1 to 3 crashes over the prior 5 year period

18% had 4 or more crashes over the prior 5 year period

Objective: To identify measures of visual processing associated with crash involvement by older drivers, in a prospective follow-up study.

$Subjects received the following sensory tests:

Letter Acuity - ETDRS chart

Contrast Sensitivity - Pelli-Robson chart

Stereoacuity - TNO Test

Disability Glare - MCT-8000 (VisTech)

Visual Field Sensitivity - Humphrey Field Analyzer 120-point program for central 60 degree radius field

$Subjects received comprehensive eye exam resulting in a primary diagnosis (cataract, age-related maculopathy, glaucoma, diabetic retinopathy)

$Mental status was assesses using the MOMSSE, with composite score ranging from 0-28. Lower scores = higher functioning. Scores greater than 9 = cognitive impairment.

$Visual Attention was measured with the Vision Attention Analyzer:

Subtest 1: Visual processing speed (score 0-30% reduction; impaired = >0)

Subtest 2: Divided attention (score 0-30 % reduction; impaired = >14)

Subtest 3: Selective Attention (score 0-30% reduction; impaired = >28)

Overall UFOV composite score (0-90% reduction of the maximum 30 degree field size)--impaired UFOV =40% reduction or greater

$AOn the road@ exposure was estimated using questionnaire data on number of days/week subjects drove and annual number of miles driven. Subjects were asked if anyone had ever suggested they limit or stop driving.

Dependent variable: Motor vehicle crash occurrence during the 3 years following clinic assessment, obtained from Alabama Department of Public Safety. Person-years to first crash was calculated from enrollment date; Person-miles of travel was calculated by multiplying person-years times reported annual mileage.

University of Alabama, Birmingham

Ophthal-

mology clinic

$56 S=s had at least 1 crash in the 3-year follow-up period, and 11 of these had 2 or more.

$Estimated annual crash rate was 7.4 per 100 person-years of driving and 7.1 per million person-miles of travel.

$Crash involvement in prior 5-year period was significantly associated with increased crash risk (Risk ratio = 2.0)

$S=s who reported that someone had suggested they limit or stop driving were no more likely to be involved in a crash.

$Impaired UFOV was the only visual processing variable associated with increased crash risk.

$Significant, independent associations with crash risk in 3-year follow-up were found only for:

$UFOV reduction of > 40%: RR=2.3; 95% CI = 1.27 - 4.29)

$Driving < 7 days/week: 48% decreased crash risk (95% CI = 0.27 - 1.01).

$Dx of diabetic retinopathy (5X greater risk, 95% CI = 1.13 - 21.8).

$Dx of glaucoma: (RR=5.20, 95% CI = 1.19-22.72). Relationship for glaucoma and crashes stronger for males (RR=9.81) than females (RR=5.14).

$Of UFOV component scores, speed of processing (subtest 1) and selective attention (subtest 3) were NOT associated with crash occurrence. Impairment in divided attention (subtest 2) was significantly associated with a 2.3 fold increased risk of crashing (95% CI = 1.24 - 4.38, p=0.01).

$For every 10 points of UFOV reduction, S=s had 16% increase in crash risk.

$Estimates are that 24% of older driver crashes are due to UFOV reduction >40%.

Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998).

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

193 older drivers between age 55-87 (mean = 71 years), identified through Alabama Department of Public Safety Files.

78 drivers (cases) had at least 1 crash in the prior 5-year period that resulted in an injury to anyone in the involved vehicles.

115 drivers (controls) had no crashes in the same 5-year period.

53% male
47% female
83% White
17% African American

Objective: to identify visual risk factors for vehicle crashes by older drivers that result in injury.

$Subjects received the following sensory tests:

Letter Acuity - ETDRS chart (impairment = worse than 20/40)

Contrast Sensitivity - Pelli-Robson chart (impairment = log CS of 1.5 or worse)

Stereoacuity - TNO Test (impairment = 500 arcseconds or worse)

Disability Glare - MCT-8000 (VisTech) (impairment = values > 0)

Visual Field Sensitivity - Humphrey Field Analyzer 120-point program for central 60 degree radius field (impairment = loss of sensitivity > 1 log unit [10 dB])

$Subjects received comprehensive eye exam resulting in a primary diagnosis (cataract, age-related maculopathy, glaucoma, diabetic retinopathy)

$Mental status was assesses using the MOMSSE, with composite score ranging from 0-28. Lower scores = higher functioning. Scores greater than 9 = cognitive impairment.

$Visual Attention was measured with the Vision Attention Analyzer (composite scores from 0 - 90% reduction).

$@on-the-road@ driving exposure was estimated, using subjects= responses to a questionnaire that asked how many days per week and how many miles per year they drove.

$Presence vs absence (self-reported) of common chronic medical conditions was determined through interview questions.

Dependent measure: involvement in a crash in the previous 5-year period that resulted in an injury to anyone in the involved vehicles.

_______________________________________________

FINDINGS (Cont=d)

NOTE: although medication information was not collected in this study, Glynn et al. (1991) reported that the use of topical eye medications in elderly patients with glaucoma increased their risk of falling (an adverse mobility outcome).

University of Alabama, Birmingham

$No significant differences between cases and controls were found with respect to driving habits (exposure).

$The odds ratio for a case driver having one or more chronic diseases was 2.2 (95% CI=1.1 to 4.5).

$Case drivers were 2.1 times more likely to receive a score of greater than 9 on the MOMSSE as compared to control drivers.

$Univariate analyses showed that older drivers involved in injurious crashes were more likely to have impairments in stereoacuity (OR=2.2); visual field sensitivity (OR=2.6 for central and 2.4 for peripheral); and UFOV reductions (OR =5.3 for 23 to 40% reduction; 16.3 for 41 to 60 % reduction; and 22.0 for greater than 60% reduction).

$ Univariate analyses for common eye diseases in the elderly showed that of the 4 conditions considered, only glaucoma and macular degeneration had significantly elevated point estimates. Case drivers were 3.6 times more likely to report a diagnosis of glaucoma compared to controls; case drivers were 3.3 times more likely to have macular degeneration.

$ Only 2 variables were independently associated with crash risk in the multivariate analyses: UFOV and glaucoma.

$ UFOV reductions of 22.5-40%, 41-60%, and >60% were associated with 5.2, 16.5, and 21.1-fold increased risk of an injurious crash, respectively compared to those with reductions of <22.5%.

$Cases were 3.6 times more likely to report glaucoma than were controls.

Owsley, McGwin, Ball, K. (1998).

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

239 older drivers, ages 56-90 years at enrollment, drawn from the population of licensed drivers in Jefferson County over age 55.

Mean Age = 70.36 (sd = 8.95)

112 females
127 males
82% Caucasian
18% African American

S=s were recruited from the larger sample of drivers participating in the larger study (Ball et al., 1993); those with poor visual acuity were excluded (since those w/ acuity worse than 20/50 uniformly fail the first subtest of the UFOV).

Objective: to examine the utility of a set of commonly used neuropsychological tests in comparison to the UFOV in predicting state-recorded, at-fault crashes in the prior 5 year period in a group of older drivers.

Measures included:

$ MOMSSE (total performance score)

$ Trail-Making Part A (time to complete)

$ Trail Making Part B (time to complete)

$ Wechsler Memory Scale - Visual Reproduction Subtest /WMS-VR (total raw score)

$Rey-Osterrieth Complex Figures Test (accuracy scores for copy trial and immediate recall trial)

$UFOV/Visual Attention Analyzer (composite % reduction; impaired/fail > 40% vs unimpaired/pass < 40%

Crash records were obtained from the Alabama Department of Public Safety; At-fault accidents were determined by 3 independent raters.

Safe drivers = 0 at-fault crashes (n=115)

Crashers = 1 or more at-fault crashes (n=124)

________________________________________________

FINDINGS (Cont=d)

Individual Analysis of Each Cognitive Variable for Predictiveness

$Each measure significantly associated with crash status; however, sensitivity & specificity less than that achieved with UFOV reduction score.

Measure OR Sensitivity Specificity
MOMSSE 1.16 61.3 58.3
Trails A 1.02 57.3 62.6
Trails B 1.00 50.8 60.0
WMS-VR 0.90 66.1 52.2
Rey-O copy 0.95 50.0 61.7
Rey-O immed. 0.95 64.5 47.0

 

Model using UFOV pass vs fail:

$Model statistically significant (p<.0001, OR = 33.92, 95% CI = 16.54, 69.50)

$Classification success = 85.4%, with sensitivity of 86.3% and specificity of 84.3%. (same ad MODEL 3 continuous UFOV score)

University of Alabama, Birmingham

MODEL 1- predictive ability of traditional neuropsychological tests

(MOMSSE, Trails A Time, Trails B Time, WMS-VR score, Rey-O copy score & immediate recall score)

$Model statistically significant (p<.01)

$MOMSSE score & Trails A time uniquely accounted for differences in outcome of crash status.

$Classification success = 58.6% overall, correctly identifying 57.3% of crashers (sensitivity) and 60.0% of non crashers (specificity).

MODEL 2 - traditional tests & UFOV

$Model statistically significant (p<.001)

$UFOV added unique info., accounting for crash status.

$Classification success = 77.4% overall, with sensitivity of 76.6% and specificity of 78.3%.

$Model 2 significantly better than Model 1 (p<.001)

MODEL 3 - UFOV alone

$Model statistically significant (p<.001).

$Classification success = 85.4%, with sensitivity of 86.3% and specificity of 84.3%.

$No significant difference between Model 3 and Model 2.

$Estimated probability of crashing w/ UFOV score of 20 = 22%; for UFOV score of 60 = 81%.

Goode, Ball, Sloane, Roenker, Roth, Myers, and Owsley (1998).

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

53 drivers ages 57-83 (mean age = 70), recruited from the Primary Care Clinic of the School of Optometry at the University of Alabama at Birmingham.

Subjects had valid AL licenses and drove at least 1,000 mi/yr

A model 2000 Visual Attention Analyzer was used to measure the detection, localization and identification of suprathreshold targets in complex displays. The size of a person=s UFOV is determined by manipulating three variables: target presentation duration, the competing attentional demands of the central and peripheral task, and the salience of the peripheral target. Three subtests provide a measure of the percentage reduction of a maximum 35 degree radius field.

The dependent variables were the total number of state-recorded accidents in the previous 5-year period, total number of convictions for traffic violations in the past 5-year period. (This was because self-reported accidents and state-reported accidents had a very poor relationship, r=.11).

The objective of the research was to determine whether incorporating eye health (ratings of the media, central vision, peripheral vision problems, diagnosis of cataract), visual function [acuity (Bailey-Lovie Chart), contrast sensitivity (Pelli-Robson), stereoacuity (Randot, TNO, Frisby), disability glare (Vistech MCT 8000), color discrimination (Farnsworth), visual field (Humphry Visual Field analyzer)], UFOV (Visual Attention Analyzer) , and mental status (MOMSSE test: information, abstraction, digit span, orientation, verbal memory, visual memory, speech, naming, comprehension, sentence repetition, writing, reading, drawing, block design, total score) could predict number of accidents in the sample.

Subjects also completed Driving Habits Questionnaire to obtain measures of self-reported accident frequency and a composite measure of driving avoidance.

University Vision Laboratory

$UFOV was related to measures of peripheral vision (central and peripheral sensitivity loss) and to night acuity.

$Only the mental status total score and UFOV were significantly related to state-reported accidents. Only the UFOV was related to traffic citations.

$Eye health alone was not linked to accidents, and visual function alone was not significantly related to accidents.

$There were significant zero-order correlations between UFOV and accident frequency (r=.36, p<0.004) and between mental status and accidents (r=0.34, p<.02).

$S=s who failed the UFOV had 4.2 times more accidents than those who passed. S=s with high MOMSSE composite mental scores experienced 3.5 times more accidents than those with scores less than 10. Together these variables predicted accident frequency, accounting for 20% of the variance R2=0.20, f(2,49) =6.01, p<0.005.

$For intersection accidents, S=s who failed the UFOV had 15.6 times more intersection accidents than S=s who passed. S=s with high MOMSSE scores had 6.3 times more intersection accidents. Together, these variables predicted 29% of the variance in intersection accidents, R=.54, F(2,49) =9.8, p<0.001.

$For intersection accidents, MOMSSE: 25 correct rejections, 11 false alarms, 5 underpredictions (for no accident S=s) and 9 misses and 3 hits for accident-involved S=s. UFOV: 26 correct rejections, 14 false alarms, 1 miss, 11, hits.

Owsley, Ball, Sloane, Roenker, and Bruni (1991)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

294 drivers age 55-90 living in Jefferson County, AL, who were legally licensed to drive.

8 age categories =

55-59; 60-64; 65-69; 70-74; 75-79; 80-84; and 85+.

Visual Attention Analyzer. UFOV performance was summarized as a composite score ranging between 0-90, representing total percentage reduction of UFOV.

Crash data (total at-fault crashes) for the 5-year period before testing were provided by AL Dept. Of Public Safety, and defined 3 categories of crash frequency for the previous 5-year period: 0, 1-3, 4+. The 294 drivers were involved in 364-at fault crashes.

Test battery included tests described for Owsley, Ball, Sloane, Roenker, and Bruni (1991), plus the following cognitive tests assessing visuospatial abilities: Rey-Osterreith test; Trailmaking test; and the WAIS block design test.

_______________________________________________

FINDINGS (Cont=d)

As a predictor, UFOV resulted in 142 hits, 18 misses, 25 false-positives, and 109 correct rejections. Of the 25 false-positives, 19 were S=s who reported avoiding driving in general, avoided driving alone, and/or avoided left turns, thus minimizing their driving exposure. Removing these people from the data set increases the correlation between UFOV and crash frequency from r=0.52 to r=0.62.

UFOV had high sensitivity (89%) and high specificity (81%); mental status had sensitivity and specificity values of 61% and 62%, respectively.

University Vision Laboratory

The following corrlelations were found between crash frequency and:

Eye health =.23 (p<0.01);

Central vision* = -.24 (p<0.01);

Peripheral vision = .26 (p<0.01);

Mental status = .34 (p<0.01);

UFOV = .52 (p<0.01).

*neg correlation, because higher numbers for centrral vision = better performance; for other variables, lower numbers = better performance)

$LISREL modeling program was used to evaluate IV in terms of whether they directly influence a DV, or operate indirectly through other IV=s.

Central vision, peripheral vision, and eye health were intercorrelated; they have indirect effects on crash frequency, but direct effects on UFOV. UFOV and mental status were the only variables that had a direct effect on crash frequency, accounting for 28% of the variance in crash frequency.

$Central and peripheral vision accounted for 30% of the UFOV variance. Mental status had a significant direct effect on UFOV and crash frequency, but its effect on crash frequency was indirect, because removal of its direct effect from the model only sightly reduced the crash frequency accounted for (28% to 27%). If UFOV is entirely removed from the model, the remaining visual variables jointly account for 5% of the crash frequency. Adding the mental status to the eye status variables accounts for 16% of the variance in crash frequency. Inclusion of UFOV maximizes the prediction of crash frequency.

Ball, Owsley, Sloane, Roenker, and Bruni (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

3,669 randomly-selected Class C license renewal applicants, licensed in California for at least 12 years, and unable to renew by mail. Four driver age groups were studied:

26-39, 40-51, 52-69, and 70+.

A model 2000 Visual Attention Analyzer was used to measure the detection, localization and identification of suprathreshold targets in complex displays. Three subtests provide a measure of the percentage reduction of a maximum 35 degree radius field. The three measures were: total UFOV loss, perceptual-reaction time, and UFOV loss associated with divided attention.

5 experimental vision tests were employed:

$ Pelli-Robson Low-Contrast Acuity Test (measures loss in low contrast acuity; ability to see objects and borders)

$ Smith-Kettlewell Low-Luminance Card (measures high-contrast near-acuity loss and low-contrast near-acuity loss)

$ Berkeley Glare Tester (measures low-contrast near acuity loss, and low-contrast near-acuity loss in the presence of glare)

$ Modified Synemen Perimeter (measures standard visual field-integrity loss and attentional visual field-integrity loss

$ Visual Attention Analyzer (measures loss in UFOV, the area of the visual field in which useful information can be rapidly extracted from a complex visual display)

The dependent measure was the crash frequency during the previous 3-year period, extracted from the DMV database.

Drivers also completed a Driving Habits Survey measuring level of restriction (never, sometimes, often or always) for night driving, rain or fog, sunrise or sunset, driving alone, left turns, and heavy traffic.

_______________________________________________

FINDINGS (Cont=d)

$Of drivers age 52-69, less than 7% failed the UFOV test.

$Approximately 7% of the variation in reported level of self-restriction was explained by total UFOV and divided attention test performance (or age), and 5% for PRT (or age)--the worse the visual performance or the older the driver, the more restriction. UFOV subtests were the only measure associated with the avoidance of heavy traffic. Also avoided by 70+ age S=s with poor UFOV: total amount of driving, driving in rain and fog, avoiding parallel parking, driving alone, driving at sunrise or sunset, and making left turns.

California DMV Field Offices:

Carmichael

El Cerrito

Roseville

$S=s rated clarity of instructions high, but safety-relatedness and fairness of requirement to pass were rated lower than sensory tests. Regression analyses showed that S=s who performed more poorly on attentional tests tended to rate them more negatively.

$For all age groups combined, test score was not significantly associated with total prior 3-year crash involvement when considered in isolation.

$S=s aged 70+ showed high variability in visual divided attention ability and PRT. There was a very small percentage of drivers age 70+ with very good total UFOV.

$Test scores had small but significant predictive value (2.9%) for S=s age 70+.

$After adjusting for gender, age, and exposure, total UFOV scores explained 0.9% of the variance in crash involvement, PRT explained 0.9% and divided attention explained 0.9%.

$Association with crashes for S=s in the 70+ age group was even stronger, with total UFOV accounting for 4.1% of the variation in crashes, PRT accounting for 4.1% of the crashes, and divided attention accounting for 4.3% of the crashes in the oldest age group. UFOV not predictive of crashes in the 3 younger age groups.

$Of 285 S=s age 70+, 84 (29%) scored poorly. 36 of the 285 S=s had an accident, and of the 36, 13 (36%) scored poorly on the UFOV. Thus UFOV sensitivity = 36%, specificity=71%, positive predictive accuracy=15.5%. For citation occurrence, sensitivity=28%, specificity=70%, positive predictive accuracy=12%.

Hennessy (1995)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

1,475 ITT Hartford Insurance Co. policyholders for whom past driving histories were available through insurance records, divided into two groups based on the presence or absence of recent at-fault accidents. Driver age ranged between 50 and 80+ and was distributed as follows:

$ 26 percent of the sample were between 50-64,

$ 54 percent were between 65-74,

$ 20 percent were over 75.

Participants were active drivers who had (generally) been pre-screened for risk in the insurance underwriting process. Also, participants who came in for testing appeared confident in their driving abilities.

The Visual Attention Analyzer was employed; the overall score from the three subtests--speed of information processing, divided attention, and a measure of distractibility--was used to describe useful field of view loss.

Insurance and motor vehicle department records provided information about the following variables: at-fault accidents, non-fault accidents, non-accident claims, violations and convictions, miles driven, age, gender and marital status.

Testing rooms in hotels in 15 cities throughout Connecticut, Florida, and Illinois

Results showed that 42 percent of the sample had an at-fault accident between 1989-1991. Univariate correlations and multiple regression analyses were computed to determine the relationships between the variables and accidents.

The correlation between performance on the UFOV test and at-fault accidents (r=0.05) was significant (p<0.05).

The low correlation was explained by the possibility that because participants were recruited through their insurance company (as opposed to being recruited through an eye clinic and offered a detailed eye exam, as were the subjects in the Ball et al. [1991] study), drivers who were less confident in their driving skills may have elected not to participate for fear that their insurance rates could be affected.

Brown, Greaney, Mitchel, and Lee (1993)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

101 licensed drivers (39 females and 62 males) age 72-90 (mean age = 78.3) who were members of a preexisting study cohort engaged in longitudinal studies of a community-dwelling cohort of older people (at Buck Center for Research in Aging)

Only Part 1 of this test was used in the study. Part 1 tests processing speed for stimuli in the fovea, rather than visual field. The subject must identify a silhouette rapidly flashed in the central part of the field as either a car or a truck. Stimulus duration ranges from 16 to 500 ms; the briefest stimulus duration at which a subject could make the identification correctly 75% of the time is his/her score.

An on-road driving exam was given by the project driving instructor (owner/operator of a driving school in San Francisco) based on the California Driving Performance Evaluation (DPE), and using the same scoresheet as used for the MDPE given in San Jose by these researchers. (See On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium). A weighted error score was calculated as total # of unweighted errors, plus twice the sum of critical and hazardous errors. Concentration errors were also noted.

Critical errors = errors which would in normal circumstances cause test termination (turning from improper lane, dangerous maneuver, examiner intervention needed).

Hazardous errors = dangerous maneuver or examiner intervention.

Concentration errors = subject unable to proceed to field office at end of test, or drove past the street on which the field office was located and did not recognize their error.

Novato, Marin County California; Buck Center for Research in Aging

The correlation between performance on Module 1 of the UFOV (here called APRT@) and road test performance was moderately high, but did not reach significance (r= 0.27).

A multiple regression model using 98 subjects using age, average time per exercise on WayPoint 1, Perceptual Response Time (Part 1 of the UFOV), and average number of cognitive domains on the MMSE in which subjects made 1 error yielded a significant prediction of weighted error score on the drive test (Multiple R = .484, adjusted R2 =0.202.

Substituting channel capacity for WayPoint average time reduced the number of subjects to 92, and yielded a multiple R of .475 and adjusted R2 of 0.190.

Eliminating age and using Waypoint average time, MMSE error areas and PRT yielded a multiple R of .462, and adjusted R2 of 0.188. Substituting channel capacity for average WayPoint time yielded Multiple R of .451, adjusted R2 of 0.176.

Using only WayPoint 1 average time and PRT as predictors of weighted error score on the road test yielded multiple R = .428; adjusted R2=0.166.

PRT was (after WayPoint) the best predictor of weighted error score, with a unique contribution of R2 of 3%, after adjustment for age.

Janke and Hersch (1997)

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

Recruitment is focused on older drivers over the age of 80 in the attempt to identify a higher percentage of drivers who qualify for the training protocol.

In Progress Study: AUFOV Training Intervention.@

Subjects are screened on visual acuity, contrast sensitivity, and UFOV. Those who demonstrate impaired driving and reduced UFOV are assigned to training or no training group. Driving performance for trained group is compared to that of untrained participants.

 

Washington University

Preliminary results indicate that the average UFOV reduction for those who pass the driving evaluation is 26% while the average reduction for marginal drivers is 52%. Retraining has been complete on 1 older driver, however this individual remained marginal due to difficulty in following directions and general confusion.

An additional 20 participants are scheduled to participate in 1996-1997.

Study Funded by NIH through Roybal Center

Principal Investigator: Linda Hunt, O.T.

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

Older drivers in the Chicago area are recruited through GEICO insurance records, based on past driving record to ensure inclusion of individuals with a range of crash involvement.

In Progress Study: AUFOV Training Intervention.@

Data collection includes UFOV, visual acuity, contrast sensitivity, and visual sensitivity. Subjects are randomly assigned to the training or no training groups; crash involvement is tracked subsequent to training, along with mobility and continued driving histories.

Rehab Institute of Chicago

Recruitment underway, with approximately 700 Chicago drivers ages 55+ with points on their record relative to approximately 6,500 similar drivers with 0 points on their record.

(Center update 1/9/96)

Data collection continues through 1996-1997.

Study Funded by NIH through Roybal Center:

Principal Investigator: Christie Rom, O.T.

ATTENTION/

PERCEPTION/

COGNITION

Attentional Visual Field:

Visual Attention Analyzer (UFOV)

Active mature drivers age 48-94 (mean age = 69) who were residents of Warren county, KY and surrounding area.

Recruited through letters, telephone conversations, and public addresses to community groups. Also, letters sent to drivers age 55+ who had been involved in at least 1 crash between 1988 and 1993

During the 1st 2 years of study (1993-1995) 456 older drivers were screened for attentional difficulties with UFOV. 129 s=s with restriction of 35% or more in attentional field were identified, and 71 completed the training study. Individuals with < 30% UFOV reduction were recruited as control S=s (n=25)

Visual acuity and contrast sensitivity were assessed to ensure that poor visual attentional performance was not caused by poor visual function

S=s were divided into 1 of 2 training groups:

(1) UFOV: (n=49 subjects); four 1-hour blocks on UFOV, customized to needs of individual (processing speed, divided attention training, and/or selective attention training). Size of UFOV assessed; training continued until a mastery level of 75% correct performance was achieved (average training time = 4.5 hours).

(2) Doron Driving Simulator. (n=22 subjects) : Two educational sessions of 2 hours each. Included 3 hours of instruction in driver safety + 1 hour on-the-road demonstration of these driving skills (e.g., safe following dist, use of turn signals).

Control group (n=25): Individuals with < 30% UFOV reduction

Participants were assessed on several visual, attentional, and driving tasks; then training proceeded, and S=s were re-assessed on the same measures. These included:

$UFOV

$Simple RT to simulated brake lights (Doron L-225 Driving Simulator)

$Complex RT to Doron simulator stimuli

$15-mi open road driving evaluation (1-mile warm up, plus 2 loops of a 7-mile urban/suburban route)

Driving evaluation: Two independent evaluators in the back seat rated each driver on a checklist of 455 driving skills. Behaviors rated on 3 point scale: 0=very unsafe or inappropriate; 1=somewhat unsafe; 2=safe or appropriate. Also a global rating of driving skill was indicated, ranging from 1 (drive aborted/very unsafe) to 6 (very competent driver). Eleven composite behaviors were formed from the 455 individual items: (1) acceleration; (2) gap selection; (3) position in traffic; (4) signals; (5) speed; (6) stop position; (7) deceleration; (8) tracking; (9) turning; (10) right of way; and (11) changing lanes. A visual search composite had to be dropped from analyses due to difficulty in assessing behavior.

A dangerous maneuver composite was created from 17 high traffic roadways, consisting of 6 left unprotected turns, 9 entrances to high traffic road from stop sign, and 2 opportunities for inappropriate stopping in traffic to turn right.

Bowling Green, KY:

Laboratory and on-road, in-traffic Evaluations

$UFOV scores significantly improved across testing sessions for only the UFOV-trained subjects (average = 24.44 point improvement).

$No significant differences were found across testing sessions for Simple Reaction Time.

$For Complex RT, only the UFOV-trained group significantly improved their scores (average improvement = 0.287 sec, or 23 feet)

$On the on-road driving evaluation, both the Simulator and UFOV-trained group improved their global ratings across test sessions; there was no change in the control groups= global rating.

$For turning (turning into the correct lane), stop position (positioning vehicle at stops in order to see clearly but not obstructing traffic flow) and signals (signaling 100-150 ft in advance of a turn) composites, only the simulator-trained group significantly improved from pre- to post-training test.

$No group by pre/post interactions were found for the other composites, but general improvement was found for all groups from pre to post test. This reflects comfort and familiarity on second drive through the route.

$For the dangerous maneuvers composite, only the UFOV-trained group demonstrated a significant reduction in the number of dangerous maneuvers from pre to post test.

$Simulator training was effective in some areas of specific instruction and demonstration; UFOV training did not transfer to driving skills that reflect the mechanical operation of the vehicle, but improved items that measured critical search and judgment abilities in visually cluttered and cognitively demanding situations.

Roenker, Cissell, and Ball (Submitted)

ATTENTION/

PERCEPTION/

COGNITION

Divided Attention:

MultiCAD

82 Areferred@ subjects aged 60-91 (26 of which were identified as probably being cognitively impaired to some degree). The drivers were referred to the DMV for reexamination due to a medical condition (by physician, optometrist, ophthalmologist), a series of licensing test failures, a flagrant driving error (police referral), or some other indicator of driving impairment.

This test used MultiCAD to measure drivers' ability to remain vigilant and respond in a timely and appropriate manner to events that occurred directly ahead, in the travel path, while detecting unexpected events of a safety-critical nature that occur in the areas of peripheral vision. After angular motion sensitivity data were obtained, the same driving video continued to use the lead vehicle target as a "foveal task" (i.e., located centrally along the driver's line of sight). At predetermined intervals in relation to a (lead vehicle) brake light stimulus, vehicles and pedestrians, offset at angles of 15 degrees and 30 degrees to the left and right sides, were introduced unexpectedly in the periphery of the driver's forward vision. The motion of these peripheral targets brought them into potential conflict with the driver within several seconds' travel time at current speeds.

For threats intersecting from the periphery at approximately a 15-degree angle of eccentricity (2 trials), the measures of effectiveness were (1) mean reaction time for correct response to (a) a vehicle pulling out from behind a building on the right side of the scene and (b) a vehicle backing out of a parking space from behind a (blocking) U-Haul van on the left side of the scene; and (2) percent error for these two trials.

For threats intersecting from the periphery at approximately a 30-degree angle of eccentricity (1 trial), the measures of effectiveness were (1) mean reaction time for correct response to a pedestrian stepping of the curb and entering the driver=s path; and (2) percent error.

Multiple linear regressions were conducted to arrive at the best linear combination of variables for predicting performance on road tests (see On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium), and comparisons were made between cognitively impaired and cognitively non-impaired referral drivers to determine whether there were differences in performance on nondriving tests and driving tests.

California DMV Field Office

The proportion of errors on trials with a threat at 15 degrees was significantly correlated with weighted error score on the drive test (r=.2430, p<.043).

Neither mean time (to 15 or 30 degree targets) nor proportion of errors to threats at 30 degrees were correlated significantly with weighted error score on the drive test.

Although the cognitively impaired referrals had higher error proportions for threats at 15 and 30 degrees (did not brake in 34% of either trial type) than did cognitively unimpaired referrals (who did not brake in 16% and 30% of the trials), the differences were not significant.

Response time to targets at 15 and 30 degrees did not discriminate between cognitively impaired referrals and cognitively unimpaired referrals.

Janke & Eberhard (1998).

Staplin, Gish, Decina, Lococo, and McKnight (in press)

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Rules of the Road

105 drivers licensed in Nebraska, aged 65-88 (mean age = 71.4). 54 were females (mean age = 70.5 years); 51 were males (mean age = 72.2 years). All subjects were volunteers, and were paid $25.00 for participating. 36 had taken a driver education course in the past 10 years.

50-question, multiple choice test designed to determine the driving knowledge pertinent to the types of accidents in which older drivers in Nebraska were over-involved.

Questions pinpointed contributing circumstances (failure to yield, disregard signal, improper turn signal, improper turn, following too close, and improper lane change) and accident type (right angle, rear end, side swipe, head on, left turn, other turn, right turn, and pedestrian).

The percentage of the questions answered quickly was used as the measure of driving knowledge.

The driving performance of the subjects was evaluated using the on-street driving performance measurement (DPM) technique developed by Vanosdall and Rudisill (1979). The subjects were evaluated by a driver education expert trained in the use of the DPM technique, while they drove in their own cars. The DPM route was a 19-km circuit designed to evaluate the subjects in the situations that are most often involved in the accidents of older drivers. Therefore, their performance was evaluated at 7 intersections where they were required to make left turns at 5 intersections and right turns at the other 2 intersections. Four of the left turns were made from left-turn lanes onto four-lane divided arterial streets in suburban areas, and one was made from a left turn lane onto a two-lane one-way street in an outlying business district.

Paper and pencil test: University laboratory.

Driving measures:

business district and residential street networks

The driving knowledge test score was significantly correlated with driving performance (correlation coefficient =0.27, p=0.0053). Better performance on the knowledge test was associated with better on-road driving performance.

Of interest is the finding that whether or not a subject had taken a driver education course within the past 10 years had a very small correlation with on-road driving performance. Most of the subjects who had taken the course had taken it more than 5 years ago; therefore, the findings are not applicable to older drivers who have taken the course in the past 5 years.

Tarawneh, McCoy, Bishu, and Ballard (1993).

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Rules of the Road

$ 102 Areferred@ subjects aged 60-91 (34 of which were identified as probably being cognitively impaired to some degree). 47% of the noncognitively impaired referred drivers had visual impairment noted on their record, and 24% of the cognitively impaired had a visual disability noted). The drivers were referred to the DMV for reexamination due to a medical condition (by physician, optometrist, ophthalmologist), a series of licensing test failures, a flagrant driving error (police referral), or some other indicator of driving impairment.

$ 33 paid Avolunteers@ aged 56-85, recruited through signs posted at study site or word of mouth.

12-item multiple-choice written test with 4 alternatives per item. Items selected from DMV=s standard renewal knowledge test.

Three tiers of analyses were conducted in this research: (1) logistic regressions to determine what combination of tests, observations, or survey variables, with what weightings, would best predict whether a subject was a volunteer or referral; (2) multiple linear regressions were conducted to arrive at the best linear combination of variables for predicting performance on road tests; and (3) comparisons were made between cognitively impaired and cognitively non-impaired referral drivers to determine whether there were differences in performance on nondriving tests and driving tests.

(See On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium).

California DMV Field Office

Referral drivers made significantly more errors than did the volunteer group. The correlation between knowledge error and group was not significant, however (r= .234)

Average number of errors :

referrals = 2.70

Volunteers = 1.58

The correlation between knowledge test errors and weighted errors on the road test was significant for the combined referral and volunteer group (r=.3847, p<.000) and for the referral group only (r=.3316, p<.001).

A multiple linear regression model using knowledge test score, Auto Trails time, Doron Cue Recognition 2 score, MultiCAD Static Contrast Sensitivity time with the high contrast 20/80 target, and MultiCAD Static Acuity time for correct responses at 20/80 accounted for 56.4% of the variance in performance on the road test (weighted road test error score).

Although the cognitively impaired group had more knowledge test errors (average = 3.76) than the cognitively unimpaired group (average = 2.14), the difference was not significant.

Janke & Eberhard (1998)

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Rules of the Road

17 subjects (age 57-97; mean age = 75); 6 females and 11 males.

8 S=s were referred from local mental disorder clinics or from local physicians because of possible dementia and associated driving problems.

9 S=s were community residents who did not have suspected dementia or driving problems.

The road (driving) knowledge test was a multiple choice paper and pencil test consisting of 21 questions assessing knowledge of rules of the road. It additionally required subjects to identify and describe the meaning of 16 road signs (what the required driver action was); later, they were required to identify 6 of these signs by shape, when presented without wording, and in a black-and-white format.

An on-road driving assessment was performed with the subject driving with a certified driving examiner in a dual-brake vehicle. Simple maneuvers were first performed in a parking lot, then subjects joined the flow of traffic and traveled over a prescribed route in moderate to heavy traffic. Subjects were scored on the basis of errors or omissions that correspond to points on the State of New York road test exam; higher scores indicate poorer performance. Therefore a total score was used as well as a determination of whether the subject met or exceeded state standards ("pass") or failed to meet standards ("fail"). In addition, a pass/fail rating was given for the subjects' performance in steering control, braking, acceleration, judgment in traffic, observation skills, and turning skills (particularly left turning).

Clinical tests: University Laboratory

On-road driving evaluation: parking lot and in-traffic (moderate to heavy traffic situations)

The group that failed the road exam had significantly lower mean scores on the written (multiple choice) knowledge test and the road sign identification test.

Of the 8 persons referred for possible dementia, 5 failed the road test, 2 passed the test, and 1 was unable to complete the evaluation.

Cushman (1992)

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Traffic Sign Recognition

$ 102 Areferred@ subjects aged 60-91 (34 of which were identified as probably being cognitively impaired to some degree). 47% of the noncognitively impaired referred drivers had visual impairment noted on their record, and 24% of the cognitively impaired had a visual disability noted). The drivers were referred to the DMV for reexamination due to a medical condition (by physician, optometrist, ophthalmologist), a series of licensing test failures, a flagrant driving error (police referral), or some other indicator of driving impairment.

$ 33 paid Avolunteers@ aged 56-85, recruited through signs posted at study site or word of mouth.

Two-part written traffic-sign test. Part 1 presented pictures of traffic signs and asked whether it meant that the driver should perform a certain action (e.g., Awatch for hazards@). Part 2 presented several traffic sign shapes embedded in complex abstract drawings, and subject were to indicate the number of sign shapes of a particular type hidden in the drawing.

Three tiers of analyses were conducted in this research: (1) logistic regressions to determine what combination of tests, observations, or survey variables, with what weightings, would best predict whether a subject was a volunteer or referral; (2) multiple linear regressions were conducted to arrive at the best linear combination of variables for predicting performance on road tests; and (3) comparisons were made between cognitively impaired and cognitively non-impaired referral drivers to determine whether there were differences in performance on nondriving tests and driving tests.

(See On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium).

California DMV Field Office

The correlation between traffic sign errors and group (volunteer vs referral) was not significant.

Although the referral group made more errors on the traffic sign test (average = 8.20) than did the volunteers (average = 6.90), the difference was not significant.

Sign test errors correlated significantly with weighted errors on the road test (r=.2026, p< .044) for the combined referral and volunteer group, but not for the referral group only (r=.1046, p<.396)

Traffic sign error score did not discriminate between cognitively impaired referral subjects (average error score = 8.67) and cognitively unimpaired referral subjects (average error score = 7.96)

Janke & Eberhard (1998)

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Traffic Sign Recognition

$ Healthy elderly controls (n=13); mean age = 73.5; CDR score =0

$ Subjects with very mild dementia (n=12) ; mean age = 72.5; CDR score = 0.5

$ Subjects with mild dementia (n=13); mean age = 73.4; CDR score = 1.0

Subjects came from the Washington University Longitudinal Studies population

Dementia severity measured w/ Washington University=s Clinical Dementia Rating

Traffic sign recognition required the identification of the following four standard symbols: traffic merging, no right turn, no left turn, and no U turn. These symbol signs were chosen because they are frequently encountered in everyday driving situations. Subjects were asked to explain the meaning of each symbol. Each item was scored individually to determine if one type of sign posed greater difficulty than the others.

The in-vehicle, on-road driving ability of participants was scored independently by a driving instructor (blind to study design and dementia status of the subjects), and an unblinded occupational therapist (Principal Investigator). The vehicle was a standard model car w/ automatic transmission and equipped with dual brake pedals. Each subject drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt Apass/fail@ rating was given by each observer in the vehicle.

Washington University Alzheimer=s Disease Research Center.

$Five subjects--all in the CDR 1 stage--@failed@ the in-car on-road test. There was 100% agreement between the driving instructor and principal investigator in their pass/fail ratings for all 38 drivers. The ability to follow the driving instructor=s directions, the demonstration of appropriate decision-making (>judgment=) in traffic, and interpretation of traffic signs were highly correlated with overall driving performance.

$Traffic sign interpretation during the road test was scored as Aunsafe@ when the subject required verbal cues or physical assistance to comply with the sign=s intent. Three prohibitive signs (no right turn, no left turn, no u turn) were tested. Ability to interpret an Aactive@ sign (merging traffic) correlated significantly (p<0.01) with three Aprohibitive@ signs as follows: no right turn = 0.553; no left turn = 0.402; no u turn = 0.621.

$All 5 CDR 1 subjects who failed the road test performed poorly on the pre-driving traffic sign recognition test.

$The correlation between the pass/fail outcome on the road test and performance on the Traffic Sign Recognition test was significant at the p<.0002 level.

$The authors noted that visual form detection may be impaired in mild senile dementia of the Alzheimer type (SDAT), while visual acuity remains intact; this may contribute to the difficulty some subjects experienced with sign recognition, since the signs were symbols (form) rather than letters (acuity). There was no association between acuity and driving performance in this study.

Hunt, Morris, Edwards, and Wilson (1993)

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Traffic Sign Recognition

3,238 drivers ages 65+, who applied for renewal of North Carolina driver=s license

Test administered by license examiner, requiring the driver to identify and explain the meaning of 12 traffic signs based on their color and shape (e.g., yellow diamond with + would be identified as a warning sign for a crossroad ahead). The signs are displayed six at a time in the viewing equipment used for vision testing. The test is not normally timed for license renewal, however, for the research, examiners recorded how long (seconds) it took license applicants to complete test. Applicants were not told they were being timed; number of errors remained the only criteria for passing or failing test. Three or more errors automatically dismisses a license applicant.

Dependent variable: involvement in a police-reported motor vehicle crash during the three-year period immediately preceding license renewal.

 

Eight NC driver=s license offices, representing a mix of urban and rural locations in the western, central, and eastern portions of the State.

Performance declined significantly as a function of increasing age (time to complete test increased with increasing age).

Correlational coefficient with number of crashes = 0.05 (p<0.001). Annual crash involvements increased with increasing (poorer) cognitive scores.

Stutts, Stewart, and Martell (1996)

ATTENTION/

PERCEPTION/

COGNITION

Driving Knowledge:

Traffic Sign Recognition

101 licensed drivers (39 females and 62 males) age 72-90 (mean age = 78.3) who were members of a preexisting study cohort engaged in longitudinal studies of a community-dwelling cohort of older people (at Buck Center for Research in Aging)

Paper-and-pencil test consisting of 12 factually oriented questions requiring a subject to check an alternative corresponding to the meaning of each pictured sign, and one judgmentally oriented question, where an intersection displays a Ano left turn@ and two Ado not enter@ signs on the through path, and the subject must check the alternative corresponding to what they could do (turn right).

An on-road driving exam was given by the project driving instructor (owner/operator of a driving school in San Francisco) based on the California Driving Performance Evaluation (DPE), and using the same scoresheet as used for the MDPE given in San Jose by these researchers. (See On-road Performance Measures of Driving Safety: California MDPE at the end of this Compendium). A weighted error score was calculated as total # of unweighted errors, plus twice the sum of critical and hazardous errors. Concentration errors were also noted.

Critical errors = errors which would in normal circumstances cause test termination (turning from improper lane, dangerous maneuver, examiner intervention needed).

Hazardous errors = dangerous maneuver or examiner intervention.

Concentration errors = subject unable to proceed to field office at end of test, or drove past the street on which the field office was located and did not recognize their error.

Novato, Marin County California; Buck Center for Research in Aging

The correlation between traffic sign errors and weighted error score on the drive test was not significant (r=0.07)

Janke and Hersch (1997)

ATTENTION/

PERCEPTION/

COGNITION

Immediate/Delayed Recall:

Logical Memory Subscale of Wechsler Memory Scale

$ Healthy elderly controls (n=13); mean age = 73.5; CDR score =0

$ Subjects with very mild dementia (n=12) ; mean age = 72.5; CDR score = 0.5

$ Subjects with mild dementia (n=13); mean age = 73.4; CDR score = 1.0

Subjects came from the Washington University Longitudinal Studies population

Dementia severity measured w/ Washington University=s Clinical Dementia Rating

The Logical Memory subscale of Wechsler Memory Scale was given to participants prior to the on-road drive test. Logical memory assesses immediate or delayed recall of verbal ideas presented in two paragraphs, read aloud by the experimenter.

The in-vehicle, on-road driving ability of participants was scored independently by a driving instructor (blind to study design and dementia status of the subjects), and an unblinded occupational therapist. The vehicle was a standard model car w/ automatic transmission and equipped with dual brake pedals. Each subject drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt Apass/fail@ rating was given by each observer in the vehicle.

Washington University Alzheimer=s Disease Research Center.

Five subjects--all in the CDR 1 stage--@failed@ the in-car on-road test. There was 100% agreement between the driving instructor and principal investigator in their pass/fail ratings for all 38 drivers. The ability to follow the driving instructor=s directions, the demonstration of appropriate decision-making (>judgment=) in traffic, and interpretation of traffic signs were highly correlated with overall driving performance. Other behaviors demonstrated by subjects who Afailed@ the in-car exam included coasting to a near stop in the midst of traffic, drifting into other lanes of traffic, stopping abruptly without cause, simultaneously pressing the brake and accelerator while driving, delay in changing lanes when an obstacle appeared, and failure to understand why other drivers signaled them in frustration or exaggeration.

The correlation between the pass/fail outcome on the road test and performance on the Logical Memory test was significant at the p<.0009 level.

Hunt, Morris, Edwards, and Wilson (1993)

ATTENTION/

PERCEPTION/

COGNITION

Language Abilities/Naming Behavior:

Boston Naming Test

$ Healthy elderly controls (n=13); mean age = 73.5; CDR score =0

$ Subjects with very mild dementia (n=12) ; mean age = 72.5; CDR score = 0.5

$ Subjects with mild dementia (n=13); mean age = 73.4; CDR score = 1.0

Subjects came from the Washington University Longitudinal Studies population

Dementia severity measured w/ Washington University=s Clinical Dementia Rating

The Boston Naming Test (Kaplan, Goodglass, and Weintraub, 1976) is a test of knowledge and language abilities (Janke, 1994). Sixty line drawings representing common to rare objects are presented individually to a subject, who must name the object. The subjects were also given a word fluency test (Thurston and Thurston, 1949) and an aphasia battery.

The in-vehicle, on-road driving ability of participants was scored independently by a driving instructor (blind to study design and dementia status of the subjects), and an unblinded occupational therapist. The vehicle was a standard model car w/ automatic transmission and equipped with dual brake pedals. Each subject drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt Apass/fail@ rating was given by each observer in the vehicle.

Washington University Alzheimer=s Disease Research Center.

Five subjects--all in the CDR 1 stage--@failed@ the in-car on-road test. There was 100% agreement between the driving instructor and principal investigator in their pass/fail ratings for all 38 drivers. The ability to follow the driving instructor=s directions, the demonstration of appropriate decision-making (>judgment=) in traffic, and interpretation of traffic signs were highly correlated with overall driving performance. Other behaviors demonstrated by subjects who Afailed@ the in-car exam included coasting to a near stop in the midst of traffic, drifting into other lanes of traffic, stopping abruptly without cause, simultaneously pressing the brake and accelerator while driving, delay in changing lanes when an obstacle appeared, and failure to understand why other drivers signaled them in frustration or exaggeration.

The correlation between the pass/fail outcome on the road test and performance on the Boston Naming Test was significant at the p<.003 level. The correlation between driving performance and word fluency was not significant. Performance on the aphasia battery correlated significantly with driving performance (p<.0001). The authors note that road test performance depended, in part, on the ability to follow verbal commands. Language impairment in SDAT may interfere w/ the ability to understand commands or advice from other passengers, rendering copilots ineffective in ensuring or extending driving competency in demented drivers.

Hunt, Morris, Edwards, and Wilson (1993)

ATTENTION/

PERCEPTION/

COGNITION

Mental Status:

Mattis Organic Mental Status Syndrome Examination (MOMSSE)

53 drivers ages 57-83 (mean age = 70), recruited from the Primary Care Clinic of the School of Optometry at the University of Alabama at Birmingham.

Subjects had valid AL licenses and drove at least 1,000 mi/yr

A brief mental status examination (Mattis, 1976) consisting of items testing:

$ -General fund of information (e.g., How many weeks are in a year?)

$ Verbal Abstraction (e.g., How are a poem and statue alike?)

$ Attention (forward and backward digit span)

$ Memory (orientation, verbal memory, reproduction of design from memory)

$ Language (e.g., test for objects, body parts, double and triple commands, reading silently and aloud)

$ Construction (draw a clock, cube copying)

It is comprised of a sample of several WAIS subtests, a Benton geometric figure, and some items from the Eisenson Test of Aphasia. It requires 15 to 20 mins. to administer. Each of the 14 subtests was scored from 0 (normal) to 2 (impaired), and an overall composite score was calculated by adding subtest scores. Composite scores ranged from 0 to 28 (0 = excellent mental status; 28 = severe dementia)

Subjects were also assessed with several other measures to find predictors of accidents. Assessments included:

1. Eye health
2. Visual Function:
$ Static Acuity (Bailey-Lovie chart)
$ Contrast Sensitivity (Pelli-Robson)
$ Disability Glare (MCT 8000)
$ Stereopsis (Randot, TNQ, Frisby)
$ Color Discrimination (Farnsworth Dichotomous Test Panel)
$ Visual Field Sensitivity (Humphrey Visual Field Analyzer)
$ Useful Field of View- Visual Attention Analyzer
3. Driving Habits Questionnaire (Sloane et al., 1990) to measure self-imposed driving restrictions and self-reported accident frequency

Accident information was obtained on all subjects from the Alabama Department of Public Safety. Data obtained for each subject included total number of accidents in the last five years and the total number of convictions for violations of traffic laws.

University of Alabama at Birmingham

Mental Status (Score on MOMSSE) found to be related to number of accidents (r=.36.)

When accidents were categorized by type, most were found to be intersection problems. MOMSSE found to be better predictors of intersection accidents than accidents in genera ( r=.41). MOMSSE and UFOV together predicted 29% of the variance in intersection accidents, and 20% of the variance in accidents in general.

Individuals with high MOMSSE scores (n=8) experienced 3.8 times more accidents on average than those with MOMSSE scores < 10 (n=45).

For intersection accidents only, subjects with MOMSSE scores > 10 (n=8) had a total of 9 intersection accidents, and those with scores < 10 (n=39) had only 7 intersection accidents between them. On the basis of the number of subjects in each group, individuals with higher MOMSSE scores had 6.3 times more intersection accidents than those with lower scores.

Eye health and measures of visual function were unrelated to accidents, in and of themselves, although they contributed to UFOV performance.

MOMSSE as a predictor of intersection accidents:

$ 34 s=s were predicted to have no accidents. 25 S=s had no accidents on record, but 9 S=s did

$ 19 S=s were predicted to have 1 or more accidents. 11 had no accidents; of the 8 who had accidents, 5 had fewer than predicted by MOMSSE.

Owsley, Ball, Sloane, Roenker, and Bruni (1991)

ATTENTION/

PERCEPTION/

COGNITION

Mental Status:

Mattis Organic Mental Status Syndrome Examination (MOMSSE)

294 subjects 56-90 yrs old mean age 71 yrs.

33% 0 crashes
49% 1-3 crashes
18% 4+ crashes

Subjects evenly distributed within 7 age groups within each crash category

Age Groups: 55-59, 60-64, 65-69, 70-74, 75-79, 80-84, 85+

The objective of the study was to test a model designed to predict crash frequency in older drivers on the basis of visual and cognitive measures.

Mental status was assessed by the Mattis Organic Mental Status Syndrome Examination.

The tests described in Owsley et al (1991) were also administered to assess visual sensory function, UFOV, driving habits, and eye health.

The dependent variable was the total number of at-fault crashes recorded by the state during the 5-year period prior to testing.

 

UAB Laboratory

Significant correlation between MOMSSE score and crash frequency (r=.34, p<.01)

Significant correlation between UFOV and crash frequency (r=.52, p<.01)

Significant correlation between eye health (central vision problems, peripheral vision problems, ocular media problems) and crash frequency

Significant correlation between central vision and crash frequency (r=-.24, p<.01). Correlation is neg. Because central vision expressed in terms of log contrast sensitivity, where higher numbers represent better performance)

Significant correlation between peripheral vision and crash frequency (r=.26, p<.01)

Data tested with the LISREL VII structural modeling program to evaluate IV's in terms of whether they directly influence DV, or if they operate indirectly through other variables.

UFOV and mental status were the only variables that had a direct effect on the crash-frequency variance. Mental status was found to have a small, but significant direct effect on crash frequency, and a larger indirect effect on crash frequency through UFOV. Together, UFOV and mental status (MOMSSE) account for 28% of the variance in crash frequency.

Mental status had sensitivity (.61) and specificity (.62) values that were Amarkedly@ less than those for UFOV (.89) and (.81), respectively.

Ball, Owsley, Sloane, Roenker, and Bruni (1993)

ATTENTION/

PERCEPTION/

COGNITION

Mental Status:

Mini-Mental State Evaluation (MMSE)

105 drivers licensed in Nebraska, aged 65-88 (mean age = 71.4). 54 were females (mean age = 70.5 years); 51 were males (mean age = 72.2 years). All subjects were volunteers, and were paid $25.00 for participating. 36 had taken a driver education course in the past 10 years.

Screening instrument for dementia (Folstein, Folstein, and McHugh, 1975) that contains tests of orientation, immediate and delayed recall, backward spelling, object naming, repetition of a phrase, following a three-stage command, sentence reading and comprehension, sentence writing, and design copying. Scores range from 0 to 30.

The driving performance of the subjects was evaluated using the on-street driving performance measurement (DPM) technique developed by Vanosdall and Rudisill (1979). The subjects were evaluated by a driver education expert trained in the use of the DPM technique, while they drove in their own cars. The DPM route was a 19-km circuit designed to evaluate the subjects in the situations that are most often involved in the accidents of older drivers. Therefore, their performance was evaluated at 7 intersections where they were required to make left turns at 5 intersections and right turns at the other 2 intersections. Four of the left turns were made from left-turn lanes onto four-lane divided arterial streets in suburban areas, and one was made from a left turn lane onto a two-lane one-way street in an outlying business district.

 

Cognitive measures: University laboratory.

Driving measures:

business district and residential street networks

MMSE showed a significant correlation to performance on the driving task (correlation = 0.24, p<0.01).

Tarawneh, McCoy, Bishu, and Ballard (1993).

ATTENTION/

PERCEPTION/

COGNITION

Mental Status:

Mini-Mental State Evaluation (MMSE)

(brief form)

492 subjects age 60+

half were applicants (community group) to the state-home-based long-term care program; half were nursing home residents

An abbreviated test was constructed including the first 4 items of the MMSE in which a cut-off score of 14 provided good sensitivity and specificity, to economically identify cognitive impairment. {NOTE: this study did not address driving performance}

The brief MMSE included: (1) orientation to time, (2) orientation to place, (3) memorizing and repeating three nonrelated items (house, bus, dog), and (4) spelling Aworld@ backward.

Community-care setting

A score of 14 had a sensitivity of 98% (i.e., 2% of persons likely to be cognitively impaired would fail to have the full MMSE completed--false negatives) and a specificity of 87% (i.e., 13% of the applicants considered cognitively intact would have to complete the full MMSE--false positives). The 14-point cut-off, when applied to the nursing home population produced a sensitivity of 100% and a specificity of 82.4%.

Paveza, Cohen, Blaser, and Hagopian (1990)