Characteristics of the learning curve in robotic thoracic surgery in an emerging country

Abstract

It is not established which factors impact the learning curve (LC) in robotic thoracic surgery (RTS), especially in emerging countries. The aim of this study is to analyze LC in RTS in Brazil and identify factors that can accelerate LC. We selected the first cases of two Brazilian surgeons who started their LC. We used CUSUM and the Lowess technique to measure LC for each surgeon and Poisson regression to assess factors associated with shorter console time (CT). 58 patients were operated by each surgeon and included in the analysis. Surgeries performed were different: Surgeon I (SI) performed 54 lobectomies (93.11%), whereas Surgeon II (SII) had a varied mix of cases. SI was proctored in his first 10 cases (17.24%), while SII in his first 41 cases (70.68%). The mean interval between surgeries was 8 days for SI and 16 days for SII. There were differences in the LC phases of the two surgeons, mainly regarding complications and conversions. There was shorter CT by 30% in the presence of a proctor, and by 20% with the Da Vinci Xi. Mix of cases did not seem to contribute to faster LC. Higher frequency between surgeries seems to be associated with a faster curve. Presence of proctor and use of bolder technologies reduced console time. We wonder if in phase 3 it is necessary to keep a proctor on complex cases to avoid serious complications. More studies are necessary to understand which factors impact the LC.

Keywords: Learning curve; Learning curve in robotic surgery; Robotic surgery; Thoracic robotic surgery; Thoracic surgery.