The reliability of self-reported childhood otitis media by adults

Objective: To estimate reliability of retrospective questionnaire data on childhood recurrent otitis media, and to identify factors influencing inconsistency in self-report. Methods: Retrospective questionnaire data from a population-based Sample of 4430 Norwegian twins who participated in two questionnaire surveys, 6 years apart, containing identical questions on otitis media. Main outcome measure was,individual consistency in reporting of otitis media. The tetrachoric correlation and Cohen's Kappa were used to measure reliability. Inconsistent and consistent responders were compared on medical history and mental distress. The use of twin data made it possible to test whether inconsistent responders represent a group with intermediate levels of otitis media severity. Results: The test-retest tetrachoric correlation was 0.82 and Kappa was 0.53. Inconsistency in response was not associated with mental distress, but related to reported number of otitis media episodes per year, use of medical services and history of ear surgery. Conclusion: Retrospective self-reported otitis media is a relatively reliable measure. The study suggests that reporting inconsistency is likely to be associated with less severe disease. The finding that reporting pattern is related to disease severity is in accordance with the proposed need for a uniform agreement on the definition of the term "recurrent others media" for research purposes. (c) 2005 Etsevier Ireland Ltd. All rights reserved.