There is a need for an easily administered instrument which can be applied to all patients with restless legs syndrome (RLS) to measure disease severity for clinical assessment, research, or therapeutic trials. The pathophysiology of RLS is not clear and no objective measure so far devised can apply to all patients or accurately reflect severity. Moreover, RLS is primarily a subjective disorder. Therefore, a subjective scale is at present the optimal instrument to meet this need. Twenty centers from six countries participated in an initial reliability and validation study of a rating scale for the severity of RLS designed by the International RLS study group (IRLSSG). A ten-question scale was developed on the basis of repeated expert evaluation of potential items. This scale, the IRLSSG rating scale (IRLS), was administered to 196 RLS patients, most on some medication, and 209 control subjects. The IRLS was found to have high levels of internal consistency, inter-examiner reliability, test-retest reliability over a 2-4 week period, and convergent validity. It also demonstrated criterion validity when tested against the current criterion of a clinical global impression and readily discriminated patient from control groups. The scale was dominated by a single severity factor that explained at least 59% of the pooled item variance. This scale meets performance criteria for a brief, patient completed instrument that can be used to assess RLS severity for purposes of clinical assessment, research, or therapeutic trials. It supports a finding that RLS is a relatively uniform disorder in which the severity of the basic symptoms is strongly related to their impact on the patient's life. In future studies, the IRLS should be tested against objective measures of RLS severity and its sensitivity should be studied as RLS severity is systematically manipulated by therapeutic interventions.