Association between Glaucoma and At-fault Motor Vehicle Collision Involvement among Older Drivers: A Population-based Study

Abstract

Objective: To examine the association between glaucoma and motor vehicle collision (MVC) involvement among older drivers, including the role of visual field impairment that may underlie any association found.

Design: A retrospective, population-based study.

Participants: A sample of 2000 licensed drivers aged ≥70 years who reside in north central Alabama.

Methods: At-fault MVC involvement over the 5 years before enrollment was obtained from state records. Three aspects of visual function were measured: habitual binocular distance visual acuity, binocular contrast sensitivity (CS), and the binocular driving visual field constructed from combining the monocular visual fields of each eye. Poisson regression was used to calculate crude and adjusted rate ratios (RRs) and 95% confidence intervals (CIs).

Main outcomes measures: At-fault MVC involvement over the 5 years before enrollment.

Results: Drivers with glaucoma (n = 206) had a 1.65 times higher MVC rate (95% CI, 1.20-2.28; P = 0.002) compared with those without glaucoma after adjusting for age, and mental status. Among those with glaucoma, drivers with severe visual field loss had higher MVC rates (RR, 2.11; 95% CI, 1.09-4.09; P = 0.027), whereas no association was found among those with impaired visual acuity and CS. When the visual field was subdivided into 6 regions (upper, lower, left, and right visual fields; horizontal and vertical meridians), we found that impairment in the left, upper, or lower visual field was associated with higher MVC rates, and an impaired left visual field showed the highest RR (3.16; P = 0.001) compared with other regions. However, no association was found in deficits in the right side or along the horizontal or vertical meridian.

Conclusions: A population-based study suggests that older drivers with glaucoma are more likely to have a history of at-fault MVC involvement than those without glaucoma. Impairment in the driving visual field in drivers with glaucoma seems to have an independent association with at-fault MVC involvement, whereas visual acuity and CS impairments do not.

Figure 1

Illustrations of the binocular visual field (a) and six subregions (b).

Figure 2

Association between binocular visual impairment and rates of at-fault crashes among drivers with glaucoma (n = 206). Rate ratios were adjusted for age, gender, mental status and the other two visual impairments. The rate ratios are indicated by the horizontal position of the solid black dots. The 95% confidence intervals (CI) around the rate ratios are indicated by the width of the horizontal lines.

Figure 3

Association between binocular visual field impairment and rates of at-fault crashes among drivers with glaucoma (n = 206). Rate ratios were adjusted for age, gender, and mental status. The rate ratios are indicated by the horizontal position of the solid black dots. The 95% confidence intervals (CI) around the rate ratios are indicated by the width of the horizontal lines.

Figure 4

Illustrations of normalized driving visual field sensitivity for four individual drivers with glaucoma. Each driving visual field map was superimposed on a driver's view of the windshield, side windows of a vehicle, and the dashboard. White solid dots represent the target locations for the driving visual field test and red crosshair indicates the driver's presumed gaze while driving. Warm colors indicate more visual field defects compared to reference normative data (see the Methods for detailed information). The panels (4a, 4c) on the left side show the visual field maps of two drivers who had a history of at-fault MVC involvement while the panels (4b, 4d) on the right side show those of drivers who were not involved in any at-fault MVC. The patients in the right column had similar age, visual acuity and contrast sensitivity whereas the degree of visual field impairment considerably differed: the ones on the left side had severe visual field impairment (within the lowest quartile); the ones on the right side did not.