To evaluate if question phrasing and patient numeracy impact estimation of urinary frequency.

We conducted a prospective study looking at reliability of a patient interview in assessing urinary frequency. Prior to completing a voiding diary, patients estimated daytime, and nighttime frequency in 3 ways: (1) how many times they urinated (2) how many hours they waited in between urinations (3) how many times they urinated over the course of 4 hours. Numeracy was assessed using the Lipkus Numeracy Scale.

Seventy-one patients completed the study. Correlation of estimates from questions 1, 2, and 3 to the diary were not statistically different. Prediction of nighttime frequency was better than daytime for all questions (correlation coefficients 0.751, 0.754, and 0.670 vs 0.596, 0.575, and 0.460). When compared to the diary, Question 1 underestimated (8.5 vs 9.7, P = .014) while Question 2 overestimated (11.8 vs 9.7, P = .027) recorded voids on a diary. All questions overpredicted nighttime frequency with 2.6, 2.9, and 3.9 predicted versus 1.6 recorded voids (P < .001). Although not statistically significant, for each question, the predicted frequency of numerate patients was more correlated to the diary than those of innumerate patients.

When compared to a voiding diary for daytime urinary frequency, asking patients how many times they urinated underestimated, and asking patients how many hours they waited between urinations overestimated the number recorded voids. Regardless of phrasing, patients overestimated nighttime urination. Patients in our functional urology population have limited numeracy, which may impact accuracy of urinary frequency estimation.