Evaluation of the Learning Curve in Robotic-Assisted Total Knee Arthroplasty: A Time-Series Analysis of Surgical Time

Abstract

Objective This study aimed to assess whether surgical times in robotic-assisted total knee arthroplasty (rTKA) could be comparable to those of conventional total knee arthroplasty (TKA), and to identify the point in the learning curve at which this occurs. Methods A time-series analysis was conducted on the first 50 consecutive primary rTKA procedures performed by a single surgeon at a university hospital. The surgeon, with seven years of experience and performing 30 TKA surgeries annually, used the ROSA® (Zimmer Biomet, Warsaw, IN, USA) robotic system for all procedures. Surgical times were analyzed using a Markov switching model to detect significant changes in operative duration. A Kruskal-Wallis test was applied to compare surgical times between rTKA and conventional TKA groups, with post hoc analysis conducted to assess differences between the first and second epochs of rTKA and conventional TKA. Results Markov analysis revealed two distinct epochs in surgical performance, with a significant reduction in surgical time occurring after the 20th case. The median surgical time in the first epoch was 118 minutes, compared to 105 minutes in the second epoch. The conventional TKA group had a median surgical time of 100 minutes. Statistical analysis, using the Kruskal-Wallis test, identified significant differences among the groups, with post hoc testing revealing no significant difference between the second epoch of rTKA and conventional TKA. Conclusions Surgical times for rTKA become comparable to those of conventional surgery after approximately 20 cases. These findings suggest that rTKA does not negatively impact operating room efficiency once the surgeon gains proficiency with the technology.