Abstract
OBJECTIVES: The objective is to compare the learning curves between two pioneer and three second-generation surgeons for RAPN in terms of WIT, CD and positive surgical margins. MATERIALS AND METHODS: The charts of consecutive RAPNs of three centres were reviewed from the UroCCR prospective database. The experience was assessed by a regression model for each group. There was a univariate analysis on three consecutive sequences of 15 procedures. The learning speed for WIT was explored graphically by polynomial regression after cubic splines. Finally, CUSUM charts were obtained. RESULTS: There were 1203 RAPN in the pioneer group and 119 performed by second-generation surgeons. There was a significant difference in the distribution of tumour size (p < 0.001) and the RENAL score (p < 0.001). The operative time was longer in the first group (p > 0.001). Independent factors for a higher WIT were the second group (p < 0.001), higher experience (p < 0.001) the collinearity between the group and experience (p < 0.001), the RENAL score (p < 0.001) and blood loss (p < 0.001). Adjusted Loess regressions showed a plateau of WIT at 400 procedures for the pioneers and a significant decrease at 20 procedures for the second generation. CUSUM chart analysis showed a 'staircase' pattern of the learning process, with three major steps at 150, 200 and 300 procedures. The major limitation is the difference in sample size between the two arms. CONCLUSIONS: Learning curve patterns would reflect a transfer of knowledge to the second-generation, as opposed to the establishment of standards by the pioneers.