Objective To assess the driving fitness of patients with glaucoma by identifying specific areas and degrees of visual field impairment that threaten safe driving.

Design Case–control study.

Setting, and participants This prospective study included 36 patients with advanced glaucoma, defined as Humphrey field analyzer (HFA; 24-2 SITA standard program) measurements of mean deviation in both eyes of worse than −12 dB, and 36 age-matched and driving exposure time-matched normal subjects. All participants underwent testing in a novel driving simulator (DS) system. Participants were recruited between September 2010 and January 2012.

Main outcome measures The number of collisions with simulated hazards and braking response time in 14 DS scenarios was recorded. Monocular HFA 24-2 test results from both eyes were merged to calculate the binocular integrated visual field (IVF). The position of the IVF subfields in which the collision-involved patients had lower sensitivity than the collision-uninvolved patients was compared with the track of the hazard. The cut-off value to predict an elevated risk of collisions was determined, as were its sensitivity and specificity, with the area under the receiver operating characteristic (AUROC) curve.

Results Patients with advanced glaucoma were involved in a significantly higher number of collisions in the DS than the age-matched and driving exposure time-matched normal subjects (119 vs 40, respectively, p<0.0001), especially in four specific DS scenarios. In these four scenarios, IVF sensitivity was significantly lower in the collision-involved patients than in the collision-uninvolved patients in subfields on or near the track of the simulated hazard (p<0.05). The subfields with the largest AUROC curve had values ranging from 0.72 to 0.91 and were located in the paracentral visual field just below the horizontal.

Conclusions Our novel DS system effectively assessed visual impairment, showing that simulators may have future potential in educating patients.