Validation of a virtual driver assessment tool for older drivers

This study aims to develop a cost-effective computer-based driving assessment tool to validate against an on-road driving assessment. The study also sought to determine: 1) relative validity of virtual test against on-road performance in healthy older adults, 2) sensitivity and specificity for identifying at-risk older drivers, 3) association with cognitive, sensory and motor skills. Future work will also examine 4) predictive validity against 24 month crash report data, and 5) the reliability and internal consistency of scenarios. The virtual driving assessment was developed by creating a virtual environment modelled on the Curtin-Woden area of Canberra where the standardised on-road driving assessment is set. Five scenarios were developed based on areas of the on-road driving route that contained challenging intersections or were identified as most discriminative of at-risk drivers. Four of the five scenarios were accompanied by audio directions (‘driver instructed’), and the final scenario required self-navigation using road signs to reach a hospital. The scenarios were piloted on a small number of older adults, and a scoring protocol developed to match the criteria and scoring methods used in the standard on-road driving test. In conclusion, a virtual driving assessment was developed that was closely matched to standard on-road driving assessment and scoring criteria. The assessment set-up is a low cost, easy to score assessment that is a valid predictor of on-road driving performance. Future work will establish its reliability with repeated testing, as well as its validity against long-term driving behaviour. The assessment can then be used as a safe and cost-effective outcome measure in studies of interventions or medication effects on driving.