Can a quantitative means be used to predict flow patterns: Agreement between visual inspection vs. flow index derived flow patterns

Abstract

INTRODUCTION: Uroflowmetry is a first-line tool in the evaluation of children with lower urinary tract symptoms. Unfortunately, there is a tremendous amount of intra- and interobserver variation in defining the shape of curves. Regrettably, one observer can see one flow as a tower and the other can call it a bell. Here lies the major flaw with the interpretation of uroflow shapes. Previously, we have shown that there is a good correlation between the calculated flow index (FI) and the shape of a flow curve (bell, tower and plateau) in normal children.

OBJECTIVE: Our hypothesis was to show that the FI-defined shapes were as good or better than the present system of grading flow curves. If so this would help remove subjectivity from the field of uroflow assessment and make studies objective and easily compared.

STUDY DESIGN: Consecutive uroflows of children who were being evaluated for lower urinary tract symptoms from two centers were reviewed and compared alongside those of presumed normal voiders; the shape of the curves were read by the same experienced readers at each institution. Only curves that were read as bell, plateau, and tower were compared with the curve patterns derived from the quantitative methods derived by one of the authors. FI was derived by taking the actual Qmax/estimated Qmax.

RESULTS: There were 591 males and 1039 females who had uroflow studies, of these 409 and 819, respectively were read as either bell, towers, or plateaus The highest kappa value for males was 0.71 (CI 0.64-0.79) using a 3 × 3 matrix indicating substantial agreement. In females the highest kappa was 0.52 (CI 0.46-0.59) in the group 2 patients using the receiver operating characteristic cutoffs but the 1 SD cutoff was close with a kappa value of 0.51, which indicates moderate agreement (Table).

DISCUSSION: Using the FI method we saw that there was substantial to moderate agreement using a quantitative method to define flow shapes based on the kappa values that were obtained in this study. The one fundamental flaw with shape determination in both visual and FI-derived methods is that the cutoffs are arbitrary since the visual defined shapes are the basis for the FI shapes.

CONCLUSION: Our findings clearly showed that a FI-derived method of defining shapes is as accurate as visual inspection or more depending on the study. The greatest disagreement occurs in those grey or transition zones between plateau, bell and tower. This system could be useful in removing much of the ambiguity and difficulty in reading flows.