Article révisé par les pairs
Résumé : Objectives: To evaluate the relative accuracy of a computerised method to quantitatively assess maximum urinary flow. Methods: A total of 1147 uroflows were evaluated by the computerised method and by three experts from different European countries. The sample consisted of uroflows from the respective visits by a 20% sample of randomly chosen patients (n = 223) with lower urinary tract symptoms with participation in two clinical trials in which the efficacy and safety of Permixon was evaluated. The proportions of automated maximum flow values included in the 10% extended range of experts (and their 95% confidence intervals) were assessed, as well as the concordance coefficients between experts and the computerised method and the paired Student's t-test for the average differences between experts and computer. Results: The rate of agreement between experts and computer varied between about 95 and 100% over factor levels for visit, type of machine and country. Concordance coefficients indicated good agreement between experts and the automated method. When looking at average differences between experts and the computer, the smallest differences were observed between experts 2, 3 and the computer (differences not statistically significant). Statistically significant average differences were observed between expert 1 and the other experts as well as between expert 1 and the computer. Conclusions: The computerised assessment decreases the fraction of variability of maximum urinary flow caused by artifacts as well as intra- and inter-expert variation. The computerised assessment of maximum urinary flow is an efficient, consistent and valid approach to quantitatively assess maximum urinary flow in clinical trials. © 2002 Elsevier Science B.V. All rights reserved.

Documents en relation

DI-fusion