Reliability of a road test after stroke

Objective: To determine the reliability of the road test performed by stroke patients. Design: Prospective study of a 6-month predriving evaluation. Setting: Driving safety center in Belgium. Participants: Thirty patients with sequelae of stroke. Interventions: Not applicable. Main Outcome Measures: Results of driving performance as judged by 2 assessors from the Center for Determination of Fitness to Drive and Car Adaptations (CARA), in a car fitted with a video camera. A third assessor also evaluated all the video recordings. Interrater reliability was evaluated by comparing results from real-life performance and video recording, as judged by the CARA assessors and video judgments between CARA assessors and the third assessor. Results: Most subitems of the road test showed more than 80% scoring agreement between the various evaluations. Intraclass correlation coefficients (ICCs) of the items varied from -.08 to 1.0. The ICC of the overall performance was .62 when real-life scores were compared with video evaluations and .80 in video versus video comparison. Conclusion: The reliability of assessing overall performance of stroke patients in the road test is moderately high and better when assessed using the same evidence. Yet, the reliability of some items needs further attention.