Background: The modified Rankin scale (mRS) is commonly employed as a measure of functional outcome after stroke. The purpose of this study was to investigate the reliability of the mRS using an expanded guidance scheme and a corresponding questionnaire on an unprecedentedly large scale. Methods: Neurologists interviewed patients with cerebral infarction on the basis of the questionnaire. These interviews were recorded on videotape. Raters (10 neurologists, 6 nurses and 4 physiotherapists) watched the videotapes of 30 patients interviewed and assessed the mRS according to the guidance scheme. Results: The agreement between the raters, as estimated from the intraclass correlation coefficient, was found to be satisfactory with values of 0.947 for neurologists and 0.963 for nurses and physiotherapists. The reproducibility was also satisfactory with values of 0.865 and 0.871, respectively. Conclusion: The data obtained suggest that our guidance scheme and questionnaire are useful for ensuring the quality of assessments made with the mRS.