This paper is a literature review on assessment of fitness to drive in older drivers with cognitive impairment. Early studies on dementia and driving generally failed to distinguish between safe and unsafe drivers on the basis of cognitive test performance. Predictive studies demonstrated that cognitively impaired persons as a group perform significantly worse than controls on both neuropsychological and driving measures. A high prevalence of cognitive impairment was found in groups of older drivers involved in traffic accidents and crashes. However, a large range in neuropsychological test scores has been found. Low to moderate correlations could be established between neuropsychological test results and on-road driving performance, making it difficult to discriminate between cognitively impaired subjects who are fit or unfit to drive. The review concludes with a discussion of methodological difficulties in the field of dementia and driving, including participant selection, the choice of neuropsychological tests, and the operationalization of driving performance. (JINS, 2000, 6, 480–490.)