To assess the test-retest reliability of a recently developed disease-specific quality of life questionnaire for evaluative studies in patients with Graves' ophthalmopathy (the GO-QOL), measuring visual functioning and psychosocial consequences of changed appearance. Ninety-three patients were included and completed the GO-QOL. Additional information on general quality of life and disease characteristics was obtained. Construct validity and internal consistency of the two subscales was determined, based on principal component analyses, Cronbach's alpha's and correlations with MOS-24, three subscales of the SIP, and clinical measures. Eighty-nine patients completed a second GO-QOL after two weeks including four additional questions about perceived changes in health status. Test-retest reliability was assessed by calculating intraclass correlation coefficients (ICCs) and limits of agreement, using several definitions of stable patients. Slight modifications from the original questionnaire were evaluated for their effect on the validity and reliability. The construct validity of the two subscales was confirmed and Cronbach's alpha's were 0.89 for visual functioning and 0.87 for appearance. The substantial ICCs found for both scales of the GO-QOL (ICCs above 0.80) reflect that the errors of measurement were relatively small, which supports the value of this questionnaire for clinical studies with relatively small sample sizes. The modification of the appearance scale improved the validity of the scale and resulted in less missing values. Following the recommendations of the joint committee of thyroid associations, we recommend the inclusion of HRQL-measures in clinical studies that evaluate treatments for patients with GO. The GO-QOL is a promising tool for this purpose. Copyright (C) 1999 Elsevier Science Inc.