VA Research and Development LOGO

Logo for the Journal of Rehab R&D
Volume 39 Number 4, July/August 2002
Pages 467  — 495


Driving performance in patients with mild to moderate glaucomatous clinical vision changes
Janet P. Szlyk, PhD; Daniel P. Taglia, MD, JD; Jennifer Paliga, BS; Deepak P. Edward, MD;
Jacob T. Wilensky, MD
Department of Research and Development, Chicago Department of Veterans Affairs Health Care System, West Side Division, Chicago, IL; Department of Ophthalmology and Visual Sciences, University of Illinois at Chicago College
of Medicine, Chicago, IL
Abstract — The purpose of this study was to determine the relationship between clinical measures of visual function and driving-related skills in patients with glaucoma who had good visual acuity in at least one eye and mild to moderate visual field loss. Methods. Twenty-five patients with glaucoma and twenty-nine age-equivalent normally sighted control subjects were included in the study. We tested each patient on an interactive driving simulator and collected vision data, including Lighthouse visual acuity, Goldmann and Humphrey visual fields, and Pelli-Robson contrast sensitivity. Information about real-world accident history for the previous 5-year period was obtained. Results. The glaucoma patients did not have significantly more simulator or real-world accidents than the normally sighted group. There were no significant differences between the groups in performance on seven of the eight simulator indexes that were measured. Of the clinical visual function measures, only lower contrast sensitivity in the eye with better contrast sensitivity correlated with driving skills, including slower speeds (r(24) = 0.58, p £ 0.01), more lane boundary crossings (r(24) = -0.54, p £ 0.01), and longer braking response times (r(24) = -0.60, p £ 0.01) for the patient group.
Conclusion. Reduced contrast sensitivity may be important in indicating the level of driving skills for individuals with glaucoma, who have normal or near-normal visual acuity and mild to moderate visual field loss.
Key words: contrast sensitivity, driving performance, glaucoma.

This material was based on work supported by a research grant from the U.S. Department of Veterans Affairs, Veterans Health Administration, Rehabilitation Research and Development Service; Chicago VA Health Care System, West Side Division, Chicago, IL; National Institutes of Health core grant EY01792; and an unrestricted grant from Research to Prevent Blindness, Inc., New York, NY.
Address all correspondence and requests for reprints to Janet Szlyk, PhD; Chicago VA Health Care System, West Side Division, 820 South Damen Avenue (M/C 151), Chicago, IL 60612; 312-996-1466; fax: 312-996-6685; email: janeszly@uic.edu.
INTRODUCTION

The standard clinical tools to diagnose and follow the progression of glaucoma are intraocular pressure measurement, optic disc examination, and perimetry. A number of visual function tests other than perimetry also show changes in glaucoma. Some studies have provided evidence that contrast sensitivity is the earliest vision function affected by glaucoma and that it correlates well with the progression of the disease [1-4]. Visual acuity, however, has limited clinical utility for diagnosing and following the progression of glaucoma. Visual acuity typically remains intact until late in the course of the disease, even when extensive paracentral and peripheral visual field loss has occurred.

The relationship between clinical visual function and general quality of life of individuals with glaucoma has only recently received analysis. Gutierrez et al. used standardized questionnaires to assess the effects of glaucoma on quality of life and performance of vision-related activities, such as driving [5]. The authors found that these standardized questionnaires are sensitive to visual field loss. Most interesting is that responses to a question rating glaucoma patients' difficulty with driving significantly correlated to the Humphrey visual field AGIS score in the patients' better-seeing eye. In another study of patients with glaucoma by Parrish et al., the authors correlated self-perceived visual task functioning as measured with the National Eye Institute Visual Functioning Questionnaire (NEI VFQ) and level of visual impairment [6]. They found that perceived driving difficulty was modestly correlated with visual field impairment.

Other studies specifically examined the relationship between visual field loss and actual driving performance, but only in defined ophthalmologic disease groups other than glaucoma, including macular degeneration and retinitis pigmentosa (RP). One study of which we are aware has examined the relationship between contrast sensitivity measurements and driving performance [7]. It found that contrast sensitivity, when combined with visual acuity and horizontal visual field data, significantly correlated to past accident involvement in drivers over age 65 years. In a recent review by Owsley and McGwin, the authors provide summaries of a number of epidemiologic studies that they say have "hinted that glaucoma may have a role in crash involvement among the elderly" [8, p. 537]. Owsley and McGwin point out that a problem with these studies was a lack of control for comorbid conditions or driving exposure. To our knowledge, however, no study has yet examined the relationship between clinical measures of visual function and actual driving performance in individuals with glaucoma.

Over the past decade, our laboratory has established the validity of evaluating driving performance using a driving simulator assessment system that we developed in collaboration with the Atari Corporation. We have demonstrated that performance on this simulator correlates to real-world accidents and driving performance, as well as to clinical measures in defined ophthalmic study groups and in age-matched controls [9-11]. In addition, we have developed predictive models for accident involvement for patients with various retinal degenerations. Establishing how visual function affects important life activities such as driving is fundamental, we believe, to validating and improving the diagnosis and treatment of ophthalmologic disease. Knowledge in this area will likely lead to improving society's approach to visual impairment, such as providing more rational vision screening criteria for licensing drivers.

In the present study, one of our aims was to determine whether mild to moderate glaucomatous damage affects driving-related skills. We did this by comparing the driving simulator performance of glaucoma patients to that of a group of age-equivalent control subjects. A second aim was to identify which clinical measures of glaucomatous damage correlate with driving perform-ance and accidents.

METHODS
Patients

Twenty-five patients who met the following criteria were included in our study:

1. Intraocular pressure in both eyes higher than 21 mm Hg (measured with Goldmann tonometry) on at least two occasions.
2. Glaucomatous visual field loss and an increased cup-to-disc ratio in at least one eye.*
3. Visual acuity of 20/50 or better in at least one eye.
4. Minimal nuclear sclerosis.
5. No age-related macular degeneration or other macular disease.
6. No diabetic retinopathy (three patients had diabetes, but no diabetic retinopathy in either eye).
7. Minimal or no posterior subcapsular cataracts.
8. Absence of physical or mental disabilities.
9. Still driving daily, as self-reported, and with a valid and unrestricted driver's license.

Patients receiving miotic therapy and those using medications that might possibly affect driving performance were excluded from the study.

The patients enrolled in the study included 14 men and 11 women, with an average age of 57.4 years (standard deviation [SD] ± 17.3). The average number of years of regular driving was 35.5 (SD ± 15.1). Three patients had diabetes mellitus, but suffered no clinically significant complications from this condition.

*All the patients had visual field loss in at least one eye with increased cup-to-disc ratio. Two patients had one eye with full fields and no disc damage in that eye; their fellow eye had moderate field loss and disc damage. Five patients had mild field loss in both eyes and disc damage in both eyes. One patient was monocular with mild field loss and disc damage in his only eye. The remaining patients had moderate visual field loss in both eyes and disc damage in both eyes.
Control Subjects

The study had 29 control subjects, 12 men and 17 women; all had normal visual acuity, normal contrast sensitivity, and normal results upon ophthalmologic examination. The control subjects' average age was 58.5 years (SD ± 18.9). The average number of years of driving was 26.9 years (SD ± 13.6). No control subjects had diabetes mellitus. We found no statistically significant differences in age (T(52) = 676.0, p £ 0.849), sex (Chi-square (1) = 0.639, p £ 0.424), or years of driving experience (t(52) = 1.76, p £ 0.088), between the control and patient groups. The Institutional Review Board approved the project protocol, and all patients and control subjects signed a consent form. All subjects were instructed that we were testing their driving skills and that all test results would be kept confidential, would be reported anonymously, and would not have deleterious consequences on their current driving privileges.

Clinical Measures of Vision

We measured visual acuity under standard conditions using the Lighthouse Distance Visual Acuity Charts and protocol described by Bailey et al. and measured contrast sensitivity using the Pelli-Robson Letter Contrast Sensitivity Charts and protocol [12-14]. All the patients in the study had visual field measurement with the equivalent of a Goldmann III-4-e target. This target was chosen over the II-4-e and V-4-e targets because of its high correlation to real-world accidents in patients with visual field loss demonstrated in previous studies and the use of the III-4-e target in defining legal blindness [10,11]. Twenty-two of the twenty-five glaucoma patients had visual fields measured with the Goldmann perimeter. Three of these patients did not have Goldmann visual fields done because a technician skilled in kinetic perimetry was not available on the day of their visit. However, these patients did have their fields measured with the Humphrey perimeter. Full-threshold perimetry data were collected from 16 patients who had routinely been followed with the Humphrey 24-2 program.

We analyzed the visual field data from a number of perspectives. We calculated the following parameters from the Goldmann visual field data:

· Total horizontal extent in degrees for each eye with the III-4-e target (determined by summing the total remaining visual field along the horizontal meridian in each eye individually).
· A sum of the total horizontal extent values across eyes.
· Total peripheral extent with the III-4-e target (determined by identifying the furthest temporal visual field locations of sensitivity in either eye from the central fixation point and then summing these values).

We calculated the following parameters from the Humphrey visual field data and the Goldmann visual field data within an approximate 24° radius:

· The percentage of points that fell below 10-dB sensitivity (the equivalent of the Goldmann III-4-e target) for each of four quadrants for each eye.
· An average percentage sensitivity loss within each quadrant averaged across both eyes' equivalent retinal areas (retinal view).
· An average percent sensitivity loss for each of four binocular visual field quadrants (visual field view).

In this way, we could gain complete information about visual field sensitivity within an approximate 24° radius from the information obtained with either the Humphrey or the Goldmann perimeters.

Driving Assessment

All the study subjects underwent testing on an interactive driving simulator that has been previously described in detail [9-11]. The simulator consisted of a seat, steering wheel, gas and brake pedals, and automatic transmission. The visual display system was composed of three 62.5-cm color monitors synchronized to display the appropriate view of a computer-generated environment. The visual display provided 160° of horizontal viewing field and 35° of vertical viewing field to the subject seated 57.5 cm from the center screen. Subjects were instructed to operate the simulator as they would normally drive their own car and to obey all traffic signs and signals along the roadway. Subjects participated in a training session on a practice course lasting approximately 15 minutes. They then drove on the actual test course for 8 minutes, during which we collected data on their performance.

Eight different indexes of the subjects' performance on the driving simulator were analyzed:

· Braking response to a stop sign (defined as the elapsed time between the presentation of a stop sign and the initiation of a brake pedal response).
· Number of lane boundary crossings (defined as any of the four tires crossing over any of the lane's boundaries).
· Horizontal eye movement (in pixels).
· Slope of the brake response curve (defined as the slope of the deceleration curve that begins with the initial brake pedal response in response to the stop sign presentation).
· Near accidents (defined as situations in which an accident is narrowly averted, as determined by an experienced observer [JPS or DPT]).
· Number of failures to stop at a stop sign or red light ("run throughs").
· Mean speed (in miles per hour).
· Number of accidents during driving simulation.
Real-World Accident and Driving Experience
Measures

All glaucoma patients and control subjects were asked to report the number of automobile accidents in which they had been involved as the driver during the last 5 years. Accidents were defined as impacts with other moving vehicles or with stationary objects that resulted in unintentional property damage or physical harm. Patients were also asked to report the number of years they had been driving regularly, defined as at least one excursion a week and at least 1,000 miles a year.

RESULTS
Data Analysis

We first analyzed for mean and median differences for the clinical vision measures for the glaucoma group and the control group using parametric t-tests and nonparametric T-tests, as appropriate. We then performed correlational analyses to compare relationships between the vision indexes for the glaucoma group. This was done, in part, to identify redundancy in the vision variables to reduce the number of correlational comparisons that would be performed with the driving indexes. Because glaucoma is not a symmetrical disease and affects several visual functions, deciding whether to select the better eye or worse eye or to attempt to combine their function and deciding whether to select a better or worse eye based on visual acuity, contrast sensitivity, or visual field extent are not straightforward decisions. For this reason, we performed these vision intercorrelations involving all the vision measures. Because of the number of statistical analyses that were performed in this study, we chose the more conservative statistical criterion for significance at p £ 0.01 to reduce the likelihood of "false" significant findings.

We next focused on the simulator and accident indexes and performed mean and median difference tests between the glaucoma and the control groups. Finally, we performed correlational analyses between the simulator and vision indexes to determine the relationships, if any, between the clinical vision measures and the simulator and accident measures.

Clinical Vision Measures

Table 1 contains the visual acuity and contrast sensitivity data for the patient group and the control group. There were no significant differences between the patients or the control subjects for visual acuity. This was the case for the eyes with the better or worse visual acuity or for the right or left eyes. However, the patients had significantly worse contrast sensitivity compared to the control subjects. This was the case for the eyes with better or worse contrast sensitivity, or for the left or right eyes.

Table 1.
Visual acuity and contrast sensitivity of glaucoma group compared to control group.
Vision Measure
Glaucoma Group
Control Group
t, p Value
 
Lighthouse Visual Acuity
Visual Acuity/Better Eye
 
 
 
Mean (±SD)
0.026 (±0.172) log MAR
-0.0235 (±0.115) log MAR
t = 1.039, 0.305
Min
-0.26 log MAR
-0.28 log MAR
-
Max
0.40 log MAR
0.18 log MAR
-
Visual Acuity/Worse Eye
 
 
 
Mean (±SD)
0.222 (±0.395) log MAR
0.0372 (±0.106) log MAR
t = 1.927, 0.061
Min
-0.18 log MAR
-0.16 log MAR
-
Max
1.6 log MAR
0.20 log MAR
-
Right Eye
 
 
 
Mean
0.088 (±0.26) log MAR
0.002 (±0.093) log MAR
t = 1.394, 0.171
Min
-0.18 log MAR
-0.18 log MAR
-
Max
1.02 log MAR
0.18 log MAR
-
Left Eye
 
 
 
Mean
0.148 (±0.26) log MAR
-0.018 (±0.132) log MAR
t = 1.385, 0.174
Min
-0.26 log MAR
-0.28 log MAR
-
Max
1.6 log MAR
0.2 log MAR
-
 
Pelli-Robson Contrast Sensitivity
Contrast Sensitivity/Better Eye
 
 
 
Mean (±SD)
1.58 (±0.171)
1.754 (±0.077)
t = -3.572, 0.001
Min
1.1
1.6
-
Max
1.8
1.85
-
Contrast Sensitivity/Worse Eye
 
 
 
Mean (±SD)
1.380 (±0.283)
1.7 (±0.092)
t = -4.087, £0.001
Min
0.8
1.55
-
Max
1.8
1.8
-
Right Eye
 
 
 
Mean
1.489 (±0.265)
1.718 (±0.082)
t = -3.122, 0.004
Min
0.8
1.6
-
Max
1.8
1.85
-
Left Eye
 
 
 
Mean
1.47 (±0.244)
1.736 (±0.095)
t = -3.872, £0.001
Min
1.05
1.55
-
Max
1.8
1.85
-

The visual field data are shown in Tables 2 to 5. Because we did not obtain visual field data for the control group for this study, we compared the visual field data for the glaucoma patients to data obtained from a control group from a previously published study in which we collected normative data for the III-4-e target on the Goldmann perimeter [15]. The control group for only the visual field measures consisted of 21 normally sighted subjects. As shown in Table 2, the glaucoma patients had significantly smaller visual field extents compared to the visual field control group. This was the case for the eyes with the better or worse horizontal extents, for the left and right eyes, for the sum of the total horizontal extents, and for the total peripheral extents. It is important to point out here that even though the glaucoma patients showed statistically significantly smaller visual fields, the minimum horizontal field extent values for at least one eye were 65° to 67°, or 105°. This was also the case for the sum of the total horizontal extent between the two eyes. The minimum total peripheral extent was 70°. These values demonstrate that the patients did have considerable visual field remaining.


Table 2.
Visual field data.
 
Visual Field
 
Vision Measure
Glaucoma Group
Control Group*
t, p Value
Total Horizontal Extent with III-4-e Target
(Goldmann Perimeter/°)
 
 
 
Horizontal Extent/Better Eye
 
 
 
Mean (±SD)
108.95 (±14.905)
127.9 (±26.6)
t = -3.17, £0.005
Min
67.0
-
-
Max
127.0
-
-
Horizontal Extent/Worse Eye
 
 
 
Mean (±SD)
81.238 (±31.148)
127.9 (±26.6)
t = -5.96, £0.0005
Min
0.0
-
-
Max
117.0
-
-
Right Eye
 
 
 
Mean (±SD)
101.476 (±16.074)
127.9 (±26.6)
t = -4.35, £0.0005
Min
65.0
-
-
Max
123.0
-
-
Left Eye
 
 
 
Mean (±SD)
88.714 (±38.048)
127.9 (±26.6)
t = -4.45, £0.0005
Min
0.0
-
-
Max
127.0
-
-
Sum of Total Horizontal Extent Values Across Eyes (°)
 
 
 
Mean (±SD)
190.286 (±40.794)
280.0 (±5.2)
t = -9.99, £0.0005
Min
105.0
-
-
Max
244.0
-
-
Total Peripheral Extent with III-4-E Target
(Goldmann Perimeter/°)
 
 
 
Mean (±SD)
129.286 (±20.558)
140.0 (±3.0)
t = -2.36, £0.025
Min
70.0
-
-
Max
140.0
-
-
*Control data from Szlyk et al. [15].
 
 
 

Table 3.
Percent visual field sensitivity loss for right and left eyes.
Vision Measure
Glaucoma Group
Right
Left
Quadrant 1
(Superior Temporal)
 
 
Mean (±SD)
13.5 (±15.6)
28.76 (±34.2)
Min
0.0
0.0
Max
62.0
100.0
Quadrant 2
(Superior Nasal)
 
 
Mean (±SD)
7.958 (±20.2)
24.28 (±39.73)
Min
0.0
0.0
Max
71.0
100.0
Quadrant 3
(Inferior Temporal)
 
 
Mean (±SD)
4.3 (±12.693)
14.1 (±29.1)
Min
0.0
0.0
Max
57.0
100.0
Quadrant 4
(Inferior Nasal)
 
 
Mean (±SD)
13.1 (±15.2)
21.7 (±29.157)
Min
0.0
0.0
Max
50.0
100.0

Table 4.
Average percent visual field sensitivity loss for each quadrant across both eyes.
Vision Measure
Glaucoma Group
Quadrant 1
(Superior Temporal)
 
Mean (±SD)
21.04 (±18.1)
Min
0.0
Max
54.0
Quadrant 2
(Superior Nasal)
 
Mean (±SD)
15.96 (±24.0)
Min
0.0
Max
85.5
Quadrant 3
(Inferior Temporal)
 
Mean (±SD)
9.1 (±17.4)
Min
0.0
Max
57.0
Quadrant 4
(Inferior Nasal)
 
Mean (±SD)
17.66 (±17.36)
Min
0.0
Max
62.50

Table 5.
Average percent sensitivity loss for each visual field.
Vision Measure
Glaucoma Group
Superior Right Visual Field
 
Mean (±SD)
19.1 (±21.8)
Min
0.0
Max
65.5
Superior Left Visual Field
 
Mean (±SD)
18.958 (±21.4)
Min
0.0
Max
78.0
Inferior Right Visual Field
 
Mean (±SD)
13.396 (±16.37)
Min
0.0
Max
50.0
Inferior Left Visual Field
 
Mean (±SD)
13.1 (±16.176)
Min
0.0
Max
50.0

Overall, the percent visual field sensitivity losses were greater for the left eye versus the right eye for the patients (Table 3). In addition, the average percent sensitivity loss for the superior left and right visual field quadrants was greater than for the inferior left and right visual field quadrants (Table 5).

Table 6 contains the correlations among the visual acuity and contrast sensitivity measures for the patients. For visual acuity, the better and worse eyes were not highly correlated (p £ 0.04), and the left and right eyes were not correlated. Because the better eye's visual acuity was correlated with the right eye's visual acuity and the worse eye's visual acuity was correlated with the left eye's visual acuity, we selected better visual acuity and worse visual acuity to use in further data presentation. Because both better and worse eye contrast sensitivity was correlated with left and right eye contrast sensitivity, we followed the pattern set with visual acuity and selected better and worse contrast sensitivity to use in further data presentation.

Table 6.
Correlation among vision indexes for glaucoma patients: Acuity versus contrast.
 
 
Acuity
 
Contrast Sensitivity
 
 
Better
Eye
Worse
Eye
Right
Eye
Left
Eye
 
Better
Eye
Worse
Eye
Right
Eye
Left
Eye
Acuity
Better Eye
-
0.409
0.706*
0.394
 
-0.726*
-0.719*
-0.773*
-0.500
 
Worse Eye
-
-
0.525†
0.888*
 
-0.305
-0.755*
-0.693*
-0.334
 
Right Eye
-
-
-
0.164
 
-0.341
-0.650*
-0.824*
-0.0955
 
Left Eye
-
-
-
-
 
-0.507
-0.720*
-0.448
-0.702*
 
Contrast
Better Eye
-
-
-
-
 
-
0.708*
0.685*
0.776*
Sensitivity
Worse Eye
-
-
-
-
 
-
-
0.842*
0.739*
 
Right Eye
-
-
-
-
 
-
-
-
0.368
 
Left Eye
-
-
-
-
 
-
-
-
-
*p £ 0.001
†p £ 0.01
 
 
 
 
 
 
 
 
 
 

Overall, visual acuity in the better eye was correlated with contrast sensitivity in the better eye. Similarly, visual acuity in the worse eye was correlated with contrast sensitivity in the worse eye.

Table 7 shows the correlations between better and worse visual acuity and contrast sensitivity and between the visual field extent measures for the patients. Neither visual acuity nor contrast sensitivity was correlated with any visual field extent measure. Most notably, the horizontal field extent in the better eye was correlated with the horizontal extent in the right eye, and the horizontal field extent in the worse eye was correlated with the horizontal extent in the left eye. Because both the better and worse horizontal field extents were correlated with the sum of the horizontal extent measure and because of their correlations with the right and left eyes, we selected better and worse horizontal field extents as the representative visual field measures to use in further data presentation. We also selected peripheral extent as a representative measure because it correlated with the sum of the horizontal extent measure and because we believed that it was important to have an index of the furthest peripheral location of sensitivity.


Table 7.
Correlation among vision indexes for glaucoma patients.
Visual Field Extents with III-4-e Target (Goldmann Perimeter)
Horizontal Extent/
Better Eye
Horizontal Extent/
Worse Eye
Horizontal Extent
Right Eye
Horizontal Extent
Left Eye
Sum of Horizontal Extent
Peripheral Extent
Acuity/Better Eye
0.228
-0.277
-0.0272
-0.148
-0.149
-0.475
Acuity/Worse Eye
0.279
-0.313
0.140
-0.230
-0.160
-0.0122
 
 
 
 
 
 
 
Contrast/Better Eye
-0.0889
0.101
0.181
-0.0562
0.0342
0.364
Contrast/Worse Eye
-0.221
-0.0723
0.0145
-0.189
-0.144
0.0939
 
 
 
 
 
 
 
Horizontal Extent/Better Eye
-
0.271
0.661*
0.356
0.592*
0.0317
Horizontal Extent/Worse Eye
-
-
0.144
0.943†
0.936†
0.806†
Horizontal Extent Right Eye
-
-
-
-0.0342
0.362
-0.0154
Horizontal Extent Left Eye
-
-
-
-
0.919†
0.742†
 
 
 
 
 
 
 
Sum of Horizontal Extent
-
-
-
-
-
0.686†
Peripheral Extent
-
-
-
-
-
-
*p £ 0.01
†p £ 0.001
 
 
 
 
 
 

Tables 8 and 9 show the correlations between visual acuity, contrast sensitivity, and the percent visual field sensitivity loss measures. Again, none of the visual field measures correlated with either visual acuity or contrast sensitivity. The binocular combinations shown in Table 9 were highly intercorrelated, whereas the monocular sensitivity losses were not correlated between the left and right eyes (Table 8).


Table 8.
Correlation among vision indexes for glaucoma patients: Percent sensitivity loss.
 
 
Right Eye
 
Left Eye
 
 
Quad 1
Quad 2
Quad 3
Quad 4
 
Quad 1
Quad 2
Quad 3
Quad 4
 
 
Sup Temp
Sup Nas
Inf Temp
Inf Nas
 
Sup Temp
Sup Nas
Inf Temp
Inf Nas
 
Acuity/Better Eye
0.0791
0.133
-0.0822
0.0705
 
-0.0428
-0.159
-0.172
-0.104
 
Acuity/Worse Eye
0.302
0.219
-0.165
-0.0222
 
0.370
0.159
0.336
0.441
 
 
 
 
 
 
 
 
 
 
 
 
Contrast/Better Eye
0.0511
-0.182
0.147
-0.0991
 
-0.351
-0.124
-0.118
-0.114
 
Contrast/Worse Eye
-0.339
-0.436
0.260
-0.196
 
-0.296
0.0258
-0.0624
-0.289
 
 
 
 
 
 
 
 
 
 
 
Right Eye
Quadrant 1
-
0.348
0.115
0.555*
 
-0.0929
0.00250
-0.0914
-0.0305
 
Quadrant 2
-
-
0.298
-0.0851
 
0.183
0.204
0.312
0.117
 
Quadrant 3
-
-
-
0.372
 
0.0632
0.440
0.286
-0.0197
 
Quadrant 4
-
-
-
-
 
0.0627
0.0525
-0.0760
0.144
 
 
 
 
 
 
 
 
 
 
 
Left Eye
Quadrant 1
-
-
-
-
 
-
0.780*
0.691*
0.722*
 
Quadrant 2
-
-
-
-
 
-
-
0.757*
0.516†
 
Quadrant 3
-
-
-
-
 
-
-
-
0.800*
 
Quadrant 4
-
-
-
-
 
-
-
-
-
*p £ 0.001
p £ 0.01
 
 
 
 
 
 
 
 
 
 

Table 9.
Correlation among vision indexes for glaucoma patients: Average percent sensitivity loss.
Vision Index
For Each Retinal Quadrant Across Both Eyes
 
For Each Binocular Visual Field
Quad 1
Quad 2
Quad 3
Quad 4
 
Superior
Superior
Inferior
Inferior
Sup Temp
Sup Nas
Inf Temp
Inf Nas
 
Right
Left
Right
Left
Acuity/Better Eye
-0.0245
-0.0973
-0.181
-0.0295
 
-0.173
-0.00419
-0.169
-0.0486
Acuity/Worse Eye
0.473
0.218
0.220
0.367
 
0.234
0.385
0.290
0.376
 
 
 
 
 
 
 
 
 
 
Contrast/Better Eye
-0.304
-0.178
-0.0289
-0.133
 
-0.0919
-0.357
-0.151
-0.0371
Contrast/Worse Eye
-0.435
-0.169
0.0598
-0.312
 
-0.100
-0.439
-0.155
-0.144
 
 
 
 
 
 
 
 
 
 
Quadrant 1
-
0.748*
0.557†
0.675*
 
0.810*
0.876*
0.711*
0.652*
Quadrant 2
-
-
0.809*
0.392
 
0.904*
0.850*
0.709*
0.628†
Quadrant 3
-
-
-
0.578*
 
0.722*
0.651*
0.900*
0.844*
Quadrant 4
-
-
-
-
 
0.516†
0.574†
0.841*
0.896*
 
 
 
 
 
 
 
 
 
 
Superior Right
-
-
-
-
 
-
0.701*
0.744*
0.592
Superior Left
-
-
-
-
 
-
-
0.655*
0.667*
 
 
 
 
 
 
 
 
 
 
Inferior Right
-
-
-
-
 
-
-
-
0.878*
Inferior Left
-
-
-
-
 
-
-
-
-
*p £ 0.001
p £ 0.01
 
 
 
 
 
 
 
 
 
Simulator Performance

Of the eight simulator indexes measured, the driving performance of the glaucoma patients significantly differed from the performance of the control subjects on only one simulator index: braking response time to stop signs (Table 10). Compared with the control group, the glaucoma patients had significantly shorter braking response times to stop signs (t(52) = 3.79, p £ 0.001).


Table 10.
Differences between glaucoma group and control group on driving simulator indexes.
Simulator Index
Glaucoma Group
Control Group
t/T, p Value
Brake Response Time to Stop Signs (s)
 
 
 
Mean (±SD)
2.193 (±1.86)
4.81 (±2.87)
t = -3.787, £0.001
Number of Lane Boundary Crossings
 
 
 
Mean (±SD)
4.08 (±11.902)
4.964 (±3.361)
t = -0.377, 0.708
Horizontal Eye Movements (Pixels)
 
 
 
Mean (±SD)
28.86 (±12.88)
26.320 (±13.48)
t = 0.618, 0.540
Slope of Brake Response Curve (Rate of Change in Speed mi/s)
 
 
 
Mean (±SD)
-4.95 (±2.57)
-4.44 (±3.33)
t = -0.618, 0.539
Number of Near Accidents
 
 
 
Median
0.00
0.50
T = 548, 0.024
25th quartile
0.00
0.00
-
75th quartile
0.00
1.00
-
Number of Run Throughs (Stop Signs/Lights)
 
 
 
Median
0.00
1.00
T = 605.5, 0.157
25th quartile
0.00
0.00
-
75th quartile
1.25
2.00
-
Speed (mi/h)
 
 
 
Mean (±SD)
26.00 (±7.38)
29.35 (±7.85)
t = -1.606, 0.114
Number of Simulator Accidents
 
 
 
Median
0.00
0.00
T = 755.5, 0.240
25th quartile
0.00
0.00
-
75th quartile
1.00
0.00
-

Table 11 contains the accident data for the patients and the control subjects. There were no significant differences for self-reported accidents, simulator accidents, or near accidents on the simulator between the two groups.


Table 11.
Differences between glaucoma group and control group for accidents and near accidents.
Simulator Index
Glaucoma Group
Control Group
T, p Value
Self-Reported Accidents
 
 
 
Median
0.00
0.00
T = 342, 0.768
25th quartile
0.00
0.00
-
75th quartile
0.25
1.00
-
Simulator Accidents
 
 
 
Median
0.00
0.00
T = 755.5, 0.240
25th quartile
0.00
0.00
-
75th quartile
1.00
0.00
-
Near Accidents (Simulator)
 
 
 
Median
0.00
0.50
T = 548, 0.024
25th quartile
0.00
0.00
-
75th quartile
0.00
1.00
-
Correlation of Clinical Measures of Vision to
Simulator Performance

Table 12 shows the correlations between the vision measures and the driving simulator indexes for the patients. There were no correlations between any of the vision measures and the simulator indexes for the control group, so we did not include a table with these data. Also shown are the relationships between better and worse visual acuity and the simulator indexes. There were no significant correlations. Table 12 shows the relationships between better and worse contrast sensitivity and the simulator indexes. Higher (better) contrast sensitivity in the better eye was correlated with shorter braking response times, fewer lane boundary crossings, and higher speeds. There were no correlations between contrast sensitivity in the worse eye and the simulator indexes.


Table 12.
Correlation among vision indexes and driving simulator indexes for glaucoma group.
Vision Index
Visual Acuity
Brake Response
to Stop Signs
Lane Boundary Crossings
Horizontal
Eye Movement
Slope/Brake Response Curve
Run Through
Stop Signs/Lights
Speed
Better Eye
0.206
0.292
0.198
0.0774
0.0970
-0.157
Worse Eye
-0.0765
0.0378
0.410
0.100
0.227
-0.0719
Contrast Sensitivity
Better Eye
-0.604*
-0.544*
-0.373
-0.233
-0.0154
0.576*
Worse Eye
-0.347
-0.371
-0.437
-0.252
-0.0733
0.197
Visual Field
Better Eye's Horizontal
Extent
0.278
0.0436
0.184
0.343
0.0142
-0.104
Worse Eye's Horizontal
Extent
0.108
0.0361
0.0677
0.135
-0.139
-0.0639
Peripheral Extent
-0.160
-0.230
-0.0408
0.0933
-0.0413
0.0114
 
 
 
 
 
 
 
Superior Right Field
0.0483
-0.0996
0.174
-0.119
-0.136
-0.224
Superior Left Field
0.141
0.338
0.465
0.0915
-0.0485
-0.443
 
 
 
 
 
 
 
Inferior Right Field
0.236
0.0652
0.0830
-0.134
0.0140
-0.294
Inferior Left Field
0.183
-0.0223
0.137
-0.134
0.264
-0.126
*p £ 0.01
 
 
 
 
 
 

Table 12 also shows the correlations with the visual field measures and the simulator indexes. There were no significant correlations between these measures. Contrast sensitivity was the only clinical measure of vision to correlate with the simulator measures.

Correlations with Accidents

We used Pearson product moment correlational analysis to determine whether either self-reported real-world accidents or simulator accidents and simulator near accidents correlated with the clinical measures of vision or the other simulator indexes. We found no significant correlations between any of the accident measures and any of the clinical vision measures for either group. Only one relationship approached significance, and it was between self-reported real-world accidents and simulator accidents for the glaucoma patients (r(24) = 0.45, p £ 0.02).

Table 13 shows the correlations between accidents and the simulator indexes. For both the glaucoma group and the control group, more incidents of running through stop signs and traffic lights were significantly related to a higher number of near accidents on the simulator. For the glaucoma group only, a steeper braking response curve, indicative of more abrupt braking, was significantly correlated with a higher number of accidents on the simulator.


Table 13.
Correlation among driving indexes, near accidents, and accidents.
Simulator Index
Glaucoma Group
 
Control Group
Self-Reported Accidents
Near
Accidents
Simulator
Accidents
 
Self-Reported Accidents
Near
Accidents
Simulator
Accidents
Brake Response to Stop Signs
0.167
0.299
-0.249
 
0.110
0.289
-0.0157
Number of Lane Boundary
Crossings
0.119
-0.125
0.0891
 
0.123
0.0319
-0.0654
Horizontal Eye Movements
0.279
-0.288
-0.258
 
-0.0503
0.376
-0.0873
Slope of Brake Response Curve
0.00952
0.181
-0.475*
 
-0.0145
0.219
0.0165
Number of Run Throughs
(Stop Signs/Lights)
-0.00942
0.550*
0.0766
 
0.0967
0.462*
-0.258
Speed
-0.139
0.111
0.158
 
0.0797
0.0736
0.227
*p £ 0.01
 
 
 
 
 
 
 
Regression Models for Accidents

We performed multiple regression analyses on the glaucoma patients' data to determine which factors significantly predicted real-world accidents. From these analyses, we developed a model to predict accident involvement for the glaucoma patients. An equation including the sum of the data on braking response time to a stop sign, lane boundary crossings, slope of the brake response curve, and contrast sensitivity provided the best predictive model for real-world accidents (r = 0.92, variance accounted for 85 percent, p £ 0.001). Therefore, contrast sensitivity data, combined with data from the driving simulator braking and lane boundary crossing indexes, provided the best prediction of glaucoma patients' involvement in real-world accidents.

DISCUSSION

We investigated the relationship between clinical vision measures (visual acuity, contrast sensitivity, and visual fields) and driving-related skills as measured with a driving simulator in patients with glaucoma. A normally sighted control group was not statistically different for visual acuity from the glaucoma patient group (as may be seen in Table 1). Six glaucoma patients (24 percent of the glaucoma group) had worse than 20/40 in one or both eyes. Of these six patients, one had 20/800 in one eye and 20/20 in the other eye, two had 20/200 in one eye and 20/20 in the other eye, one had 20/50 in one eye and 20/32 in the other eye, one had between 20/40 and 20/50 visual acuity in each eye (0.36 log MAR [minimum angle of resolution] OD and 0.38 log MAR OS), and one had 20/50 (0.40 log MAR) visual acuity in his only eye. The patients were significantly different from the control subjects for contrast sensitivity and for visual field extent and sensitivity. For the glaucoma patients, visual acuity was correlated with contrast sensitivity; however, no correlations were found between these central vision measures and the visual field measures.

The glaucoma group was compared to an age-, sex-, and driving-experience-equivalent normally sighted control group on the eight driving indexes. The patient group was not significantly different from the control group on all driving indexes except one, braking response time to stop signs. The glaucoma patients, as a group, had shorter braking response times, which indicates a tendency to stop more rapidly upon perceiving a stop sign. The control subjects, however, would stop more gradually, reaching a point closer to the intersection. This finding could be attributed to different abilities in visual perception between the two groups or to a testing effect, where the patients may have displayed more caution during a session where they knew that their performance was being evaluated.

We also explored the possibility that this finding could have been due to faulty depth perception. Where we did not collect information on stereoscopic depth perception, we do have information about discrepant visual acuities between the two eyes that could produce deficits in depth perception. As stated previously, we had six patients who had worse than 20/40 visual acuity in at least one eye and had unequal visual acuities. We compared the braking response times of the six patients (mean = 2.38 ± 1.71 s) to the remainder of the patient group (2.13 ± 1.96 s), and we found no significant difference (p = 0.78). When the six patients were compared to the control group mean (4.81 ± 2.87), the significant difference found between the glaucoma patient group and the control group on this index was reduced to an effect with a p = 0.055. The rate of deceleration (slope of the braking response curve) was slightly higher in these six patients (-5.11 ± 1.87 mi/s) compared to either the remainder of the patient group (-4.90 ± 2.82 mi/s) or the control group (-4.44 ± 3.33 mi/s), but we found no significant statistical differences. These findings suggest that discrepant visual acuities between the two eyes, and possibly stereoacuity, were not factors that contributed to the difference between the glaucoma patients and the control subjects.

The overall conclusion from these data is that the patient group performed as well as the control subjects on the driving simulator. In addition, the findings on the simulator were related to the real-world performance of the patients and the control subjects, because there were no significant differences in the numbers of real-world accidents and the simulator accidents for the glaucoma group compared to the control group.

For the glaucoma patient group, lower (worse) contrast sensitivity in the better eye was correlated with slower speeds, more lane boundary crossings, and longer braking response times. As was just discussed, the glaucoma patients as a group had overall shorter braking response times compared to the control group, which likely may be due to behavioral hypervigilance in the patient group. However, when one considers the within-group glaucoma data alone, a significant relationship exists between better contrast sensitivity and shorter braking response times. These data actually support the hypervigilance argument in that those glaucoma patients with better contrast sensitivity could see the sign sooner and stop even faster than those with worse contrast sensitivity. Other than contrast sensitivity, we found no correlation with the driving performance of the glaucoma patients for the remaining clinical measures of vision. However, it must be stated that, in general, the patients' performance was not statistically different from that of the control subjects. These relationships account for the variance within the normal range.

These results are consistent with a previous study of driving with age-related macular degeneration (AMD) [9]. In the AMD study, we found similar relationships between contrast sensitivity and speed and between contrast sensitivity and lane boundary crossings. However, in contrast to the present study with patients with good visual acuity, in the AMD study where the patients had binocular visual acuities ranging from 20/30 to 20/100 (mean = 20/70), we found significant correlations between worse visual acuity and slower speeds, longer braking response times, and more lane boundary crossings. Consistent with the present study, the AMD patients did not have significantly more real-world accidents than a control group. The AMD patients were shown to self-restrict their driving by driving at slower speeds, not driving in unfamiliar areas, reducing nighttime driving, and taking overall fewer risks while driving (e.g., not changing lanes).

The lack of correlation between the visual field measures and driving performance appears to contradict results of several prior studies, some performed in our laboratory, and the quality of life work referenced earlier by Gutierrez et al. and Parrish et al [5,6], which have established that peripheral field loss impairs driving skills and increases the likelihood of car accidents [10,11,16-18]. However, if you consider the level of visual field impairment and the visual acuity of the glaucoma group, the results do not contradict previous work. Szlyk et al. demonstrated that patients with binocular visual field loss caused by RP had impaired driving skills and an increased likelihood of accidents [10]. These results correlated to the extent and area of their binocular horizontal visual field loss but, unlike the glaucoma patients in this study, the visual field loss of the RP patients was substantial in both eyes [10]. Consistent with these results, Johnson and Keltner reported that among a group of subjects in California who passed the state's drivers license vision screening test, those with binocular peripheral field loss had a statistically significantly greater number of accidents [17]. In addition, 35 percent of the drivers with glaucoma demonstrated a loss of visual field sensitivity; however, no information is provided about the correlation of this loss with numbers of accidents.

In light of these studies, there are several possible explanations why we did not find a correlation between visual field loss and driving performance. The inclusion criteria for this study required visual field loss in only one eye, not both eyes. Glaucomatous damage is often asymmetric, and review of the visual fields in this study revealed that many patients had normal or near-normal visual fields in their better-seeing eye. Twelve of twenty-two patients had a Goldmann horizontal field extent to a III-4-e target of 112° or greater in the better eye, and eight of sixteen patients had a mean deviation on Humphrey perimetry of -1.86 or better. Not surprisingly, the median values for the perimetry data were skewed substantially toward the upper quartile, well above the mean. Some studies have provided evidence that no significant difference exists between monocular and binocular drivers in overall driving performance, number of accidents, and level of driving skills [18-20]. Consistent with these studies, perhaps too few of our patients had sufficient visual field loss in their eye with less visual field loss to affect their driving performance. In addition, we must emphasize that the glaucoma patients' overall acuities were not significantly different from normal either in the better or the worse eye.

Furthermore, excluding our single monocular patient who had a horizontal visual field of 65°, the minimum for the sum of the total horizontal binocular visual field was 105°. In the Szlyk et al. study on driving with RP [10], we showed that 88 percent of patients with constricted binocular visual fields less than or equal to 100° diameter (III-4-e Goldmann target) had one or more accidents in the prior 5 years, whereas only 25 percent of the patients with greater than 100° diameter visual fields had one or more accidents in the same time period. A more recent study in our laboratory on clinical vision measures and their relationship to actual daily function in RP patients showed that the patients began demonstrating difficulty on tasks within the general categories of "peripheral detection" or "mobility," not specifically driving tasks, if monocular horizontal visual field extents fell below 70° in diameter (III-4-e Goldmann target) [15].

Our results indicate that contrast sensitivity may be a more sensitive visual screening tool for driving perform-ance than visual acuity or visual field measurements in this patient group with glaucoma, with normal or near-normal visual acuity, and with mild to moderate visual field loss. As mentioned earlier in this article, other studies have provided evidence that contrast sensitivity is the first visual function affected by glaucoma [1,2]. Furthermore, contrast sensitivity is easier to measure and more rapidly assessed than visual field loss. Therefore, we agree with Decina and Staplin that contrast sensitivity may be an important component in driver's license screening examinations, especially for patients with glaucoma [7].

By requiring regular driving as an inclusionary criterion for the study, we have limited our patient sample. Many, if not most, patients with moderate or severe glaucoma stop driving because of the deterioration in their driving skills. Those who continue driving are likely to have better preservation of the visual functions critical for driving. Therefore, our study actually may understate the extent to which glaucoma and the resulting deterioration in clinical measures of vision actually may affect driving performance. However, our study highlights the importance of identifying early clinical changes that may impact driving performance before the patient feels the need to self-restrict, a feeling that often occurs after an adverse experience on the road. Future research will be aimed at testing glaucoma patients with greater vision loss. We must also mention that contrast sensitivity loss, while typically seen with glaucoma, is nonspecific to this disease and can be due to macular disease and cataracts [21]. However, because the patients in this study were screened for macular disease and the presence of cataracts, we feel that the contrast sensitivity loss seen in our patients was due to glaucoma.

ACKNOWLEDGMENT

This paper was presented in part at the annual meeting of the American Academy of Ophthalmology, San Francisco, California, on October 26, 1997.

REFERENCES
1. Arden GB, Jacobsen JJ. A simple grating testing for contrast sensitivity: preliminary results indicate value in screening for glaucoma. Invest Ophthalmol Vis Sci 1978; 17:23-32.
2. Hitchings RA, Powell DJ, Arden GB, Carter RM. Contrast sensitivity gratings in glaucoma family screening. Br J Ophthalmol 1981;65:518-24.
3. Motolko MA, Phelps CD. Contrast sensitivity in asymmetric glaucoma. Int Ophthalmol 1984;7:45-49.
4. Zulauf M, Flammer J. Correction of spatial contrast sensitivity and visual fields in glaucoma. Graefes Arch Clin Exp Ophthalmol 1993;231:146-50.
5. Gutierrez P, Wilson MR, Johnson C, Gordon M, Cioffi GA, Ritch R, Sherwood M, Meng K, Mangione CM. Influence of glaucomatous field loss on health-related quality of life. Arch Ophthalmol 1997;115:777-84.
6. Parrish RK, Gedde SJ, Scott IU, Feuer WJ, Schiffman JC, Mangione CM, Montenegro-Piniella A. Visual function and quality of life among patients with glaucoma. Arch Ophthalmol 1997;115:1447-55.
7. Decina LE, Staplin L. Retrospective evaluation of alternative vision screening criteria for older and younger drivers. Accid Anal Prev 1993;25(3):267-75.
8. Owsley C, McGwin G. Vision impairment and driving. Surv Ophthalmol 1999;43(6):535-50.
9. Szlyk JP, Pizzimenti CE, Fishman GA, Kelsch R, Wetzel LC, Kagan S, Ho K. A comparison of driving in older subjects with and without age-related macular degeneration. Arch Ophthalmol 1995;113:1033-40.
10. Szlyk JP, Fishman GA, Master SP, Alexander KR. Assessment of driving performance in patients with retinitis pigmentosa. Arch Ophthalmol 1992;10:1709-13.
11. Szlyk JP, Fishman GA, Severing K, Alexander KR, Viana M. Evaluation of driving performance in patients with juvenile macular dystrophies. Arch Ophthalmol 1993;111: 207-12.
12. Bailey IL, Bullimore MA, Raasch TW, Taylor HR. Clinical grading and the effects of scaling. Invest Ophthalmol Vis Sci 1991;32:422-32.
13. Pelli DG, Robson JG, Wilkins AJ. The design of a new letter chart for measuring contrast sensitivity. Clin Vis Sci 1988;2:187-99.
14. Elliott DB, Bullimore MA, Bailey IL. Improving the reliability of the Pelli-Robson contrast sensitivity chart. Clin Vis Sci 1991;6:471-75.
15. Szlyk JP, Seiple W, Fishman GA, Alexander KR, Grover S, Mahler CL. Perceived and actual performance of daily tasks: Relationship to visual function tests in individuals with retinitis pigmentosa. Ophthalmology 2001;108(1): 65-75.
16. Fishman GA, Anderson RJ, Stinson L, Haque A. Driving performance of retinitis pigmentosa patients. Br J Ophthalmol 1981;65:122-26.
17. Johnson CA, Keltner JL. Incidence of visual field loss in 20,000 eyes and its relationship to driving performance. Arch Ophthalmol 1983;101:371-75.
18. Wood JM, Troutbeck R. Effect of restriction of the binocular visual field on driving performance. Ophthalmic Physiol Opt 1992;3:291-98.
19. Wood JM, Troutbeck R. Effect of visual impairment on driving. Hum Factors 1994;3:476-87.
20. McKnight AJ, Shinar D, Hilburn B. The visual and driving performance of monocular and binocular heavy-duty truck drivers. Accid Anal Prev 1991;4:225-37.
21. Seiple W. The clinical utility of spatial contrast sensitivity testing [chapter 114]. In: Tasman W, Jaeger E, editors. Duane's Foundations of Clinical Ophthalmology. Philadelphia (PA): Lippincott; 1992. p. 1-13.
Submitted for publication May 21, 2001. Accepted in revised form September 14, 2001.

Go to TOP

Go to the Contents of Vol. 41 No. 1