Aims: Questionnaires, including the Incontinence Impact Questionnaire (IIQ), have been used to assess the severity of urinary incontinence in clinical trials. However, the summed score derived from the IIQ does not have any inherent clinical meaning. Also, a summed score does not have interval properties, complicating the interpretation of changes in severity. The purpose of this study is to test the hypothesis that incontinence-related disability is a measurable variable and that the IIQ can measure that variable accurately.
Methods: We used Rasch analysis, a logistic regression technique, to estimate an interval scale of incontinence severity. Rasch analysis is based on the mathematical assumption that each subject's response to any given item is a function of (a) the individual's inherent level of disability and (b) the inherent difficulty of that item.
Results: Twenty-seven women with urodynamic stress urinary incontinence (USUI) completed the IIQ. Using Rasch analysis, we found that women with USUI had the most difficulty engaging in "physical recreational activities" and the least difficulty participating in "relationship with family." We also assessed the ability of the IIQ to discriminate between individuals with respect to disease severity. We found that most of the items in the IIQ are useful for discriminating incontinence severity among women with mild or moderate incontinence, but very few IIQ items distinguish among women with severe incontinence.
Conclusions: Our results have important implications for clinical trials using the IIQ: among women treated for severe incontinence, the standard IIQ summed score will underestimate the magnitude of any change in incontinence severity.
Copyright 2003 Wiley-Liss, Inc.