Test-retest reliability, validity, and sensitivity to change of the urogenital distress inventory and the incontinence impact questionnaire

Abstract

Aims: To evaluate two quality of life measures for urinary incontinence (UI) in Scottish females.

Methods: Three groups with UI from two regions in Scotland were studied. Two groups were receiving treatment for incontinence; the third was not. Women completed the UDI and IIQ twice to allow assessment of test-retest reliability and validity. Treatment groups completed the questionnaires again, postintervention, to assess ability of the measures to detect change. Other measures used to assess validity were the SF-36, HADS, weight of urine leaked, and number of incontinence episodes. By design, the three subject groups differed significantly in their characteristics, ensuring a diverse sample of women.

Results: Analysis of reliability showed a clinically trivial but statistically significant decrease in total UDI (mean, -6.1; 95% CI, -11.0 to -1.5) and IIQ (mean, -9.7; 95% CI, -15.5 to -3.9) scores between test and retest assessments, possibly due to a research effect. Most items of the UDI (18 of 19) and IIQ (28 of 30) performed very well on test-retest. The UDI and IIQ were valid in that higher scores (indicating more bothersomeness of symptoms/impact on daily living) were associated with greater severity of UI. Additionally the IIQ showed the expected associations with measures of anxiety and health status.

Conclusions: Both the UDI and IIQ detected changes in women's conditions due to intervention. The measures had good psychometric properties, including test-retest reliability, across subject groups.