Frequencies of voiding and urinary incontinence are commonly measured by a patient's recall or a diary. The recommended diary length varies from 1 to 14 days, with 7 days apparently being most common. To examine the statistical precision of these different modalities, we analyzed recall data and diary data of 74 patients with urinary frequency, incontinence, or both. Recall data on voiding and incontinence frequency were systematically higher and more variable than diary data. Longer diary length provided less variable diary data. The confidence interval of diary data was calculated by applying the normal distribution to daytime voiding frequency and the Poisson distribution to daytime incontinence frequency. For daytime voiding frequency, the 95% confidence interval was estimated to be (x - 2.65, x + 2.65) (x - 1.53, x + 1.53) (x - 1, x + 1), where x is the 1-day, 3-day, and 7-day diary mean, respectively. For daytime incontinence frequency, the confidence interval depended on both the diary length and the diary mean. It was estimated to be (0, 6.39), (1.72, 4.28), (2.36, 3.64), by using a diary mean of 3 or 1-day, 7-day, and 28-day diaries, respectively. Also, it was estimated to be (0, 1.02), (1.72, 4.28), (7.66, 12.34), when the 7-day diary mean was 0.5, 3, and 10, respectively. Studies with different samples of genuine stress incontinence (n = 37) and urge incontinence (n = 25) confirmed these results. In conclusion, we believe the 7-day diary is highly reliable for estimating voiding frequency and is a reasonable option for predicting incontinence episodes. However, the diary length should be extended in a patient with rarer events of incontinence, and it should be shortened for those who are incontinent more often or who are diagnosed with voiding frequency only.
Copyright 2002 Wiley-Liss, Inc.