Inadequate repeatability of the one-hour pad test: the need for a new incontinence outcome measure

Objective To assess the reproducibility of two one-hour pad tests performed within one week using serial ultrasound scanning to obtain identical bladder volumes, and to measure the effect of patient anxiety upon test reproducibility. Design Prospective observational study. Setting Tertiary urogynaecological unit. Sample Fifty-six incontinent women undergoing 112 pad tests. Method Two one-hour pad tests were performed with natural diuresis one week apart prior to treatment. At the second test, serial ultrasound scans were performed until bladder volume reached that of the first test, followed by identical provocation. Main outcome measures One-hour pad loss, bladder volumes (Vol(1), Vol(2)), anxiety VAS questionnaire. Results Despite serial scanning, bladder volumes differed significantly. Median volume before second pad test was 541 mi,, compared with 433 mis before first test (P < 0.001). The second pad test was also significantly larger than first (median 16g vs 4g, P = 0.017), and 13/56 (23%) women were dry on the first test but incontinent on the second. In 26 women (46%) both bladder volumes were similar, but the second pad loss was still significantly greater (median 14g vs 4g, P = 0.037). The mean difference between tests was 10g and the limits of agreement were wide (ranging from -44 to +66 S difference for the test result). Women were more anxious about leaking during the first test (Median VAS during the first test was 2.8cm, compared with 0.6cm during the second test, P = 0.008). 42.5% found the second test to be more typical. Conclusion In women with similar bladder volumes, the test-retest reliability of the one-hour pad test was judged to be clinically inadequate, as the, first and second pad test could differ by -44 to +66g. Lower anxiety levels at the second test may account for this finding. The one-hour pad test is a useful baseline measure of incontinence, but the poor repeatability suggests that is not an optimal measure of post-treatment change.