Self-reported medical history was generally accurate among Japanese workplace population

To assess the validity of self-reported medical history of several diseases among the Japanese population, and to clarify to what extent the self-reported year of diagnosis for chronic diseases is different from the physician's reports.

Subjects were 8,947 persons who responded to questions about medical history in a self-administered questionnaire. Of them, 854 subjects reported one or more medical histories and gave permission to contact their physician. The physicians were then requested to provide information on 809 subjects. Valid responses of 714 subjects were collected. We compared the self-reported medical histories with those reported by the physician.

Of 15 persons who reported myocardial infarction, 13 (87%) were confirmed. Angina pectoris was verified in eight out of the 11 (73%). The confirmation proportions of hypertension, diabetes, hyperlipidemia, and hyperuricemia were 97%, 96%, 95%, and 95%, respectively. The self-reported year of diagnosis was 1.70-2.49 years earlier than the physician-reported year for chronic diseases. Agreement between the self-reported and the physician-reported years was higher, the more recent the self-reported year was.

Self-reported medical histories were generally accurate, especially for diseases with clear diagnostic criteria. However, investigators should be aware of the errors in reporting the year of diagnosis.