Transferring Surgical Expertise: Analyzing the Learning Curve of Robotic Cardiac Surgery Operative Time Reduction When Surgeon Moves from One Experienced Center to Another

Abstract

Background: Robotically assisted cardiac surgery is performed in a team setting and is well known to be associated with learning curves. Surgeon and operative team learning curves are distinct entities, with total operative time representing the entire operative team (surgery, anesthesia, nursing, and perfusion) and cross-clamp time representing mainly the surgical team. Little is known about how a team learning curve evolves when an experienced surgeon transitions from one surgical center to another. This study investigates the dynamics of the team learning curve expressed as total operative time in the case of a surgeon with previous experience transitioning to a new team. Methods: A retrospective analysis was conducted on robotic cardiac surgeries performed by a surgeon who transitioned from one experienced surgical center to another. Operative time data were collected and categorized to assess the evolution of the learning curve. Statistical analysis, including learning curve modeling and linear regression analysis, was used to evaluate changes in total time in the operating room per case. Results: 103 cases were included in Weill Cornell Medicine (2019–2023). The median patient age was 63 years, 68% were males, 90.3% of cases were repaired for degenerative mitral valve disease, and the median body mass index was 23.87. Operative time (ORT) decreased from a median of 5.00 h [95%CI: 4.76, 6.00] in the first 30 cases to 4.83 [95%CI: 4.10, 5.27] thereafter, with the apparent curve plateauing indicative of the adaptation period to the new surgical environment (p = 0.01). Subgroup analysis among mitral cases (n = 93) showed a decrease in ORT from 5.00 [95%CI: 4.71, 5.98] in the first 26 cases to 4.83 [95%CI: 4.14, 5.30] (p = 0.045). There was no difference between the initial 30 cases and subsequent cases regarding cardiopulmonary bypass time, myocardial ischemia time, reoperation for bleeding, prolonged ventilation, reintubation, renal failure, need for an intra-aortic balloon pump, readmission to the ICU, reoperation for valvular dysfunction within 30 days, pneumonia, and deep venous thrombosis. Multivariate significant predictors of longer operative time were the first 30 cases, resection-based repairs, and MAZE as a concomitant procedure. Conclusions: Total operative time can be expected to decrease after about 30 cases when an experienced robotic surgeon moves between centers. Complications and cross-clamp times are less susceptible to a learning curve phenomenon in such a circumstance, as these depend primarily on the operating surgeon’s level of experience. Understanding these dynamics can inform the planning and management of surgical transitions, ensuring optimal patient care and continued improvement in surgical outcomes.