In order to evaluate quality-of-life (QOL) questionnaires, we studied the reliability and validity of the self-rating forms: A, comprising 16 questions; B, comprising 39 questions (Table II and III), in a special study supported by the Ministry of Health and Welfare. Sixty-four patients with inoperable lung cancer and 50 patients with chronic, non-malignant disease were entered into the study. Form A was translated and modified to Japanese culture from Schipper's Functional Living Index-Cancer, and Form B was of our own design. The internal consistency of the forms was examined by Cronbach's alpha coefficient. Both forms except for aspects of physical and social concern on form B, had alpha values greater than 0.65. A concurrent validity with performance status was shown over all aspects of concern (Pearson's correlation, r greater than 0.27, P less than 0.05) except for social (forms A and B) and psychological (form B). Analysis of variance of the QOL scores revealed significantly lower scores for the cancer patients than for the non-malignant-disease patients, except for the social aspect of concern (P less than 0.001). Factor analysis, using the principal component method, identified four major factors to account for 60 and 42% of the total variance on forms A and B, respectively. The factors were: patient's mood, anxiety over disease and treatment, relationships with other people, physical capability. In conclusion, this has been the first feasible study to show that QOL forms from the West could be used in Japan with some modifications. We are at present studying successive changes in QOL for advanced lung cancer patients in clinical trials.