RESUMEN
The on-road driving test is considered a 'gold standard' evaluation; however, its validity and reliability have not been sufficiently reviewed. This systematic review aimed to map out and synthesize literature regarding on-road driving tests using the Consensus-based Standards for the Selection of Health Measurement Instruments checklist. Cochrane Library, PubMed, CINAHL, and Web of Science databases were searched from initiation through February 2018. All articles addressing reliability or validity of on-road driving tests involving adult rehabilitation patients were included. The search output identified 513 studies and 36 articles, which were included in the review. The Washington University Road Test/Rhode Island Road Test, performance analysis of driving ability, test ride for investigating practical fitness-to-drive, and K-score demonstrated high reliability and validity in regard to the Consensus-based Standards for the Selection of Health Measurement Instruments checklist. The Washington University Road Test/Rhode Island Road Test and test ride for investigating practical fitness-to-drive were analyzed based on Classical Test Theory techniques, and performance analysis of driving ability and K-score were analyzed based on Item Response Theory techniques. The frequency of studies were Washington University Road Test/Rhode Island Road Test (n=9), Test Ride for Investigating Practical fitness-to-drive (n=8), performance analysis of driving ability (n=4), and K-score (n=1). From the viewpoint of accuracy and generalization, the Washington University Road Test/Rhode Island Road Test, test ride for investigating practical fitness-to-drive, and performance analysis of driving ability were identified as highly qualified concerning on-road driving tests. However, the ability to assess real-world driving depends on various environmental conditions.