Objective: To determine the validity of a self-reported physician diagnosis of rheumatoid arthritis (RA) in elderly women.
Methods: Women with self-reported RA were contacted and permission requested to contact their physician. Physicians were mailed a questionnaire requesting information on the subject's arthritic condition. Hand radiographs obtained at study entry were read for changes of RA.
Results: The self-reported diagnosis of RA could be confirmed in 26 (21%) individuals.
Conclusion: The positive predictive value of self-reported RA was low in this cohort. Caution needs to be exercised in the use and interpretation of self-report data concerning arthritic conditions in epidemiologic studies.