Abstract
People with hemianopia are permitted to drive (following a road test) in a small but growing number of jurisdictions. We have been studying driving with hemianopia on the road and in a driving simulator, with and without peripheral prisms that provide visual field expansion. In agreement with others, we found that in a road test some hemianopes performed well enough to be deemed safe. However, in our simulator studies, where we challenge the blind side with frequent pedestrian detection events, we find wide individual variability in detection performance (10%–90%). Furthermore, when approaching intersections, many failures to scan to the blind side were recorded. Importantly, hemianopes who head scanned more to the blind side had better detection performance. In our on-road study, hemianopes responded better to unexpected events (on the blind side) and committed fewer errors requiring tester interventions when using real peripheral prisms than sham prisms; this was true for both current and non-current drivers. Although improved detection with peripheral prisms may not be sufficient to convert all unsafe hemianopes to safe drivers, they may be a useful driving aid. We argue that road tests may not provide sufficient opportunities to evaluate detection of blind-side hazards.