Association between visual field impairment and involvement in motor vehicle collision among a sample of Japanese drivers

•Investigation into motor vehicle collision, visual field and driving behaviours.•Visual field was not statistically significant in predicting motor vehicle collision.•Insight into own at-risk behaviour and health status would affect collision risk. There is an expectation that the addition of visual field testing to driving licencing procedures may contribute to detecting unsafe drivers due to visual field (VF) impairment. However, previous research has been inconclusive as to the association between VF impairment and motor vehicle collision (MVC), and there are relatively fewer trials conducted outside the United States using a sample of general drivers compared to studies using eye clinic patients. To investigate the association between ophthalmic indicators of VF impairment and involvement in at-fault MVC in the previous five years among a sample of middle-aged or older drivers. Driving licence holders aged 40 years or over (N = 546) were administered a questionnaire survey and underwent a visual assessment at a driving licence examination centre in the suburbs of Tokyo. Of these participants, 458 were current drivers. Their response to the written questionnaire and ophthalmic indicators of VF impairment were subjected to multivariate logistic regression models to predict at-fault MVC. The analyses were done while adjusting for variables pertinent to driving habit, near-miss experience and perceived driving difficulties. We compared two different dependent variables separately: at-fault MVC based on police record and self-reports. Analysis showed that ophthalmic indicators of binocular VF impairment such as Esterman score and Integrated Visual Field was not statistically significant in explaining at-fault MVC involvement. This was true both for predicting police-registered MVC and self-reported at-fault MVC. It was indicated that history of citation, presence of medical conditions and certain driving circumstances such as frequent night-time driving would increase risk of at-fault MVC in conjunction with the individual’s perception about driving difficulties. The results of this study were consistent with previous research in that ophthalmic VF impairment may not be directly associated with MVC involvement when VF impairment is not severe. However, a combination of at-risk driving behaviour (exposure to high-risk driving situations and certain near-miss experience), deterioration in health condition, and lack of insight into such circumstances would li