OBJECTIVE: Self-report questionnaires are frequently used to obtain information in epidemiological research. However, information reported by patients are sometimes inconsistent with medical records. This study compared self-reported major rheumatologic diagnoses and co-morbid conditions with those from a medical record review. METHODS: A cross-sectional survey was conducted at two tertiary academic hospitals. All patients who visited the rheumatology department from September 2, 2009 to September 13, 2009 were enrolled in this survey. Structured patient questionnaires and medical record reviews were performed in each hospital. We evaluated agreement with kappa statistics (κ) between these two data sources for major rheumatologic diagnosis and Charlson Comorbidity Index (CCI) score. Multiple logistic regression models were used to investigate factors associated with disagreement. RESULTS: A total of 369 patients were interviewed at clinic exit. Of them, 302 patients (81.8%) were female, and the average age was 52.1 years. The agreement for major rheumatologic diagnosis between the questionnaire and patient chart was good (κ=0.763). The agreement rate for all rheumatic diseases was 81.8%; rheumatoid arthritis with 94.9%, systemic lupus erythematosus with 96.3%, and ankylosing spondylopathy with 100%. Higher educational level and longer attendance at our clinic were associated with agreement between major rheumatologic diagnoses. The agreement rate for CCI score between the data sources was 76.1%. CONCLUSION: In patients with rheumatologic diseases, the agreement for major diagnoses between self-reports and the medical record review was good, although it varied with the specific disease and patient characteristics. Comparing major rheumatologic diagnoses, the agreement rate for CCI was low.