A CUSUM analysis of operative times and complications for a surgeon initiating robot assisted pyeloplasty - a predictable decrease in operative time is possible by case 30
Noah Stern, MD, Yilong Li, BSc, Peter Wang, MD, Sumit Dave, MD.
Western University, London, ON, Canada.
Introduction: The transition from laparoscopic and open surgery to robot assisted procedures in paediatric urology leads to an increase in operative times and added health care costs. Cumulative sum (CUSUM) analysis can be used to study inflection points to detect changes in operative timing and complication rates. This analysis can be used to define a procedure’s learning curve and to monitor for unacceptable complication rates when adopting a new approach. The objectives of this study are to investigate the learning curve of a single surgeon transitioning to robot assisted pyeloplasty (RAP) in the Canadian healthcare system.
Methods: Demographic and surgical data from 50 consecutive RAP performed between 2013 and 2019 were prospectively collected. Operative time was recorded by an independent operative room personnel. The CUSUM of RAP operative times (CUSUMOT) was plotted against the number of operations (CUSUMOT . The mean operative time (OT) from each phase of CUSUMOT were compared using one-way analysis of variance (ANOVA) and post-hoc Tukey test on SPSS 26. Non-risk-adjusted cumulative observed minus expected failure chart with 80% (alert) and 95% (alarm) boundary lines was constructed using 5% acceptable and 10% unacceptable complication rates.
Results: The mean OT for RAP was 157.1 ± 39.4 min for all 50 cases. One-way ANOVA analysis showed that the 3 phases of the learning curve had significantly different mean OT (p < 0.001). Post-hoc Tukey test showed that the mean OT of Phase 1 (207.3 ± 34.7 min, the initial 12 cases), Phase 2 (161 ± 17.5 min, the middle 16 cases), and Phase 3 (127 ± 19.3 min, the last 22 cases) were all significantly different (p < 0.001) [Fig. 1]. The complication rate for RAP stabilized around the acceptable level of 5% up to case 41 before it finalized at 8% overall [Fig. 2].
Conclusion: CUSUM analysis can be used to monitor surgeon progression along the learning curve, and safety when adopting a new approach. In our study, OT showed an inflection point by case 12, and another at case 28. We propose that by case 30 a surgeon transitioning to RAP can achieve a significant decrease in OT. Complication rates remained within acceptable limits throughout, indicating that RAP can be safely adopted, even at low volume centers (8-10 RAP/year). Future studies can work to establish reference values against which surgeons can monitor competence.
Back to 2021 Abstracts