Allan Chang

A probability model for ultrasound estimation of bladder volume in the diagnosis of female urinary retention.

Objective: The aim of this study was to evaluate a probability model for ultrasound estimation of bladder volume in the diagnosis of female urinary retention, which is a common condition in the postpartum period. Methods: A total of 53 women with vaginal delivery were recruited on postpartum day 1, and their bladder volumes were first estimated by ultrasound. Immediately after the ultrasound scans the bladders were catheterized for the true bladder volumes. The ultrasound-estimated bladder volumes were then compared with the catheterized bladder volumes. The data were first tested for non-normality, and then logarithmic transformed in order to conform to normality. A probability model was built using the logarithmic-transformed data, based on the following: an arbitrary cut-off value of 150 ml for urinary retention, linear regression analysis, and paired difference analysis. Results: There was a linear relationship between the ultrasound-estimated and catheterized bladder volumes, which can be represented by the equation: log10 (Vc) = 0.2959 + 0.8853 log10 (Vu), where Vc = catheterized bladder volume, and Vu = ultrasound-estimated bladder volume. A normogram was constructed, and the probability of urinary retention presented. Conclusions: The results of this study showed that our method and statistics are logical by which the probability model of urinary retention is precise enough to support clinical decision-making.