Prediction of driving after stroke: a prospective study.

BACKGROUND The process of determining whether patients with stroke should drive again often involves off-road evaluations and road tests that usually take about 2 to 3 h to complete. OBJECTIVES This prospective study sought to identify the combination of tests that best predicts fitness to drive after stroke. The main aim was to develop a short and predictive predriving assessment battery. METHODS Sixty-eight consecutive stroke patients were studied who performed a mandatory predriving assessment at the Belgian Road Safety Institute, Brussels, within 18 months. Performance in a predriving assessment included medical examination (when needed), visual and neuropsychological evaluations, and an on-road test. Based on these assessments, a physician, psychologist, and the driving safety expert who administered the tests decided if a subject was either "fit to drive,""temporarily unfit to drive," or "unfit to drive." RESULTS Logistic regression analysis revealed a combination of visual neglect, figure of Rey, and on-road tests as the model that best predicted (R(2) = 0.73) fitness to drive after stroke. Using a discriminant function that included the 3 tests of the logistic model, the fitness to drive judgments of 59 (86.8%) subjects were correctly predicted. The sensitivity and specificity of the predictions were 79.4% and 94.1%, respectively. CONCLUSION Fitness to drive after stroke can be predicted from performance on a few road-related tests with a high degree of accuracy. However, some individuals require extended assessments and further tests.