Predicting driving ability using DriveSafe and DriveAware in people with cognitive impairments: a replication study

Background/aim: Previously developed cut-off scores for off-road assessments, DriveSafe and DriveAware, were applied to data from a new sample. Our aim was to determine whether results from previous research are replicable in a different population, to further investigate the psychometric properties of the tools.

Methods: Using a retrospective cohort design, we analysed data from DriveSafe and DriveAware gathered in three driving centres in Sydney (N = 90). We calculated sensitivity and specificity of DriveSafe and DriveAware data for predicting results of on-road testing.

Results: Sensitivity and specificity for this study were very similar to those documented previously. The lower cut-off produced specificity of 96%, identical to previous calculations. The upper cut-off score yielded sensitivity of 91% compared with 93% previously. When scores from DriveSafe and DriveAware were trichotomised (i.e. 'pass,' 'fail,' and 'requires on-road testing'), they predicted on-road performance of about half of drivers with ≥ 90% accuracy.

Conclusion: Findings will add to the body of evidence suggesting that not all drivers referred to occupational therapy driving assessors require expensive, time-consuming and potentially high risk on-road assessments.