Predicting fitness to drive using the visual recognition slide test (USyd).

OBJECTIVE The authors examined the construct and predictive validity and internal reliability of the Visual Recognition Slide Test developed at the University of Sydney (VRST-USyd). METHOD A historical cohort study using retrospective descriptive analysis of VRST-USyd scores and on-road driving performance for 838 drivers with impairments was conducted. RESULTS Rasch analysis provided evidence for the construct validity and internal reliability of the VRST-USyd. Goodness-of-fit statistics for all items were acceptable. The test had high participant and item reliability indexes and separated the participants into four groups with varying levels of skill. Using a cutoff score of 95/164, the sensitivity of the test was 81%, and the specificity was 90%. However, when coupled with clinicians' judgment of participants' awareness of their driving performance during the on-road assessment, this score improved. CONCLUSION There is evidence for reliability and construct and predictive validity of the VRST-USyd. The measurement of awareness requires further research.