Accuracy of survey questions for geriatric urinary incontinence

Risk factors, natural history, consequences, therapeutic responses and costs are all likely related to type of urinary incontinence, for example stress or urge. Yet few epidemiologic type specific data are available and only 1 study has been validated urodynamically. We compare the accuracy of a typical questionnaire used in a large epidemiologic study with the criterion standard of multichannel video urodynamic testing.

The questionnaire was administered before urodynamic testing to 132 subjects 65 years old or older, of whom 80% were women, all were mobile and none was severely demented. Responses to questionnaire items were compared to the criterion standard, singly and in combination, using a total of 4 a priori and post hoc strategies, including a computerized regression tree program.

Overall, no analytic strategy correctly classified more than 67% of patients and none accurately classified even a single type of incontinence, including stress incontinence.

Short questionnaires commonly used in epidemiologic studies correlated poorly with video urodynamic testing in incontinent older adults. Previously published information regarding prevalence of the types of incontinence should be reviewed in the light of these data.