Development and validation of a vision and night driving questionnaire

Purpose - Night-time driving difficulties are a common concern of older drivers and those with eye disease. This study aimed to develop and validate a questionnaire for assessing vision-related night driving difficulties in older drivers. Methods - Items from existing vision-related quality of life questionnaires and driving studies were used to develop a questionnaire that was completed by 283 participants who reported visual difficulties for night driving (65% female, 50 to >80 years). The questionnaire included items relating to demographic and night driving characteristics (seven items), general vision ratings (eight items), vision-related night driving difficulties (11 items), and a single open question about specific night driving difficulties. The vision-related night driving difficulty items were analysed separately using Rasch analysis to form the vision and night driving questionnaire (VND-Q). Rasch analysis assessed validity and psychometric properties of the scale. Generalised linear regression models examined associations between VND-Q scores and age, gender, amount of night driving, self-rated vision, and eye conditions. Test-retest repeatability was assessed using intra-class correlation analysis and Bland-Altman methods of agreement for a subset of 30 participants. Results - Rasch analysis indicated that a nine-item VND-Q scale was unidimensional, valid and reliable, and showed excellent discriminant ability (person separation index 3.04; person reliability 0.90). Targeting was better for those with greater self-reported night driving difficulties. Participants with self-reported bilateral eye conditions and worse self-reported general vision ratings had significantly more night driving difficulties with the VND-Q scale than individuals without eye conditions (p = 0.03) and with better general vision ratings (p < 0.001). Females reported more difficulties than males (p < 0.001) and drove shorter distances at night per week which was also associated with greater difficulties (p < 0.001). A repeatability coefficient (Rc) of 2.07 demonstrated excellent test-retest repeatability. Conclusions - The nine-item VND-Q is a unidimensional and reliable questionnaire allowing quantification of the level of visual difficulties that older drivers report at night. The development of this questionnaire is an important step in providing a reliable and validated instrument for use to guide appropriate investigations, referrals, or interventions in clinical and research settings.