Refinement of the German Day-to-Day Impact of Vaginal Aging questionnaire in perimenopausal and postmenopausal women using item response theory and classical test theory.

Abstract

We aimed to assess the structural validity of the Day-to-Day Impact of Vaginal Aging (DIVA) questionnaire using item response theory/Rasch analysis and classical test theory and refine the current version of the DIVA if necessary. Postmenopausal women reporting vaginal symptoms related to menopause participated in the study. Item characteristic curves were analyzed to see whether the response categories were functioning optimally. The assumptions of the Rasch model were tested for the whole DIVA as well as for each domain separately. Exploratory factor analyses were carried out and correlations of the single items with the DIVA domains were analyzed to identify the most-fitting items. Finally, validation analyses were carried out on the refined version. We registered 185 eligible postmenopausal women. Revising the response categories of each of the four domains led to adequate looking item characteristic curves. The whole DIVA represented a multidimensional construct, however, each of the four domains fulfilled the Rasch requirements of unidimensionality, local independence, monotonicity, and an adequate model fit. Integrating item response theory/Rasch and classical test theory, two items (item 5 and item 17) showing relevant issues were identified and removed from the refined version. In the subsequent validation, the refined DIVA showed similar validation results like its original equivalent. We created a validated refined version of the DIVA, having now three response categories instead of five. With 17 items (short-version) or rather 21 items (long-version for women with recent sexual activity), the refined DIVA is more feasible and showed several excellent measurement properties.