Determinants of driving after stroke

Objective: To identify variables that best predict a team's decision of driving ability in stroke patients from a predriving assessment. Design: Retrospective study of a 2-year predriving, evaluation. Setting: Belgian Institute for Road Safety. Participants: One hundred four patients with sequelae of first stroke. Interventions: Predriving assessments and road test. Main Outcome Measures: The suitability to resume driving based on a team decision and performance in the road test. Results: Forty-one patients (39.4%) were judged suitable, 45 (43.3%) not immediately suitable, and 18 (17.3%) not suitable to drive. Correlation coefficients and comparisons between groups revealed that most variables had significant individual relationships with the learn decision and performance on the road test. After logistic regression analysis, side of lesion, kinetic vision, visual scanning, and a road test led to the best model in predicting the learn decision (R-2 = .53). The road test was the most important determinant (R-2 = .42). Multiple regression analysis showed that the combination of acuity of left and right eyes and the figure of Rey was the best subset to predict the road test (R-2 = .28). Conclusion: The predictive accuracy of the team's decision is limited, and the road test is even lower. Inclusion of more real-road-related tests in the predriving assessment is necessary.