Robotic-arm assisted total knee arthroplasty has a learning curve of 16 cases and increased operative time of 12 min

Background Robotic-arm assisted systems are increasingly used for knee arthroplasty, however introduction of new systems can involve a learning curve. We aimed to define the learning curve in terms of operative time and component placement/sizing of a robotic system for total knee arthroplasty (TKA) in a team of experienced surgeons, and to investigate mid-term patient outcomes. Methods A total of 101 consecutive patients underwent primary robotic-arm assisted TKA by three surgeons (mean 2 year follow-up). Operative times, component placement, implant sizing and reoperations were recorded. Cumulative Summation (CUSUM) was used to analyse learning curves. Patient outcomes were compared between learning and proficiency phases. Results The learning curve was 16 cases, with a 12-min increase in operative time (P < 0.01). Once proficiency was achieved, the greatest time reductions were seen for navigation registration (P = 0.003) and bone preparation (P < 0.0001). A learning curve was found with polyethylene (PE) insert sizing (P = 0.01). No differences were found between learning and proficiency groups in terms of implant survival (100% and 97%, respectively, NS) or patient-reported outcome measures at 2 years (NS). Conclusion Introduction of a robotic-arm assisted system for TKA led to increased operative times for navigation registration and bone preparation, and a learning curve with PE insert sizing. No difference in patient outcomes between learning and proficiency groups at 2 years was found. These findings can inform surgeons' expectations when starting to use robotic-assisted systems.