The Use of Cumulative Sum Analysis to Derive Institutional and Surgeon-Specific Learning Curves for Robot-Assisted...

The cumulative sum (CUSUM) approach has been adopted to evaluate surgical competence in various contexts. The CUSUM method comprises sequential monitoring of cumulative differences from a target level in performance quality over time, allowing the detection of deviations from the target. We use the CUSUM method in this study to derive both institutional and surgeon-specific learning curves for robot-assisted radical prostatectomy (RARP). We reviewed 540 patients diagnosed with localized prostate cancer (pathologic stage T2) who underwent RARP at our institution between September 2011 and December 2017. The CUSUM method was used to construct both institutional and individual-surgeon learning curves; the outcome was assessed based on whether a positive surgical margin (PSM) was present. The target PSM rate was 20%. Of seven surgeons performing robot-assisted resections for this period, four who performed ≥60 resections were assessed separately. Of 540 patients, 74 (14%) had PSMs. The institutional CUSUM chart exhibited a downward trend for the first 54 cases and an upward trend thereafter. The CUSUM chart for the earliest adopter was similar to that for the institution; that is, learning was complete at 45 cases. Two adopters did not undergo a learning curve as they consistently performed well. The last adopter required 10 cases for initial learning. However, his CUSUM chart trend became negative at 46 cases and remained so afterward. CUSUM charts are useful for monitoring surgical quality. Long-term monitoring of the PSM rate of a surgeon who has become independent allows an attending surgeon to intervene if necessary; thus, long-term quality control is assured.