Concordance of retrospective and prospective reporting of menstrual irregularity by women in the menopausal transition.

OBJECTIVES: First, to compare retrospective self-report of irregular and skipped periods by women in the menopausal transition to information recorded concurrently on calendars; and, second, to describe how participants' definitions of irregularity and skipping affected reporting.
METHODS: The sample was 161 women who returned both complete menstrual calendars and questionnaires in any year between 1997 and 2002. Irregular and skipped periods documented on menstrual calendars were compared with self-reports. Data were z-score adjusted for repeated observations in women and analyzed using Cohen's kappa.
RESULTS: Agreement between calendar and questionnaire reporting of cycle irregularity was weak (kappa = 0.192). Participants' definitions of irregularity and skipping differed markedly from those of the researchers. Agreement about skipping was stronger overall and increased after a definition of skipping was provided to study participants (kappa pre-definition = 0.597; post-definition = 0.765). Counts of skipped cycles led to the greatest differences between researchers and participants. DISCUSSION: Accuracy of self-reports of menstrual cycle irregularity or skipping is not sufficient without explanations about what is meant by those terms. Researchers and clinicians who solicit information from women about menstrual cycle patterns need to define explicitly the phenomenon of interest to insure accurate information.