Anterior cruciate ligament reconstruction learning curve study - Comparison of the first 50 consecutive patients of five orthopaedic surgeons during a 5-year follow-up.

Background: The goal of all surgical and orthopaedic training is to ensure necessary education and surgical skills without compromising the quality of operations or patient safety. Anterior cruciate ligament reconstruction (ACLR) is a common multi-staged orthopaedic surgical procedure with a learning curve. Previous studies focus mainly on learning or the learning curve of one surgeon and tunnel placements. The aims of this study were to define the learning curve in arthroscopic ACLRs, define the number of procedures needed before the surgical "knifetime" plateaus, examine the effect of experience on complications, and identify possible individual differences in the surgical learning curve.

Methods: The study included the first 50 consecutive ACLR operations of five orthopaedic surgeons, thus, a total of 250 patients. For comparison and statistical analysis, patients were arranged into five groups, each comprising 50 patients (=order group). Order group 1 comprised the first 10 patients operated on by each of the five surgeons, group 2 patients 11-20, group 3 patients 21-30, group 4 patients 31-40, and group 5 the last 10 patients. The learning curve was defined with a LOESS curve. Surgical time and complications, including graft failure and postoperative knee instability, were analysed between order groups and between surgeons.

Results: Median surgical time was 105 (interquartile range 82-124) min. The learning curve showed the first steep decline in surgical time and started to settle slowly after 20 reconstructions. Surgical time was significantly longer when order group 1 was compared with order group 2 (p = 0.031), and when order group 1 was compared separately with order groups 3, 4, and 5 (p < 0.001). Operation order alone explained only 17.1% of the alteration in surgical time. No significant difference emerged in graft failure rate between the order groups or the surgeons. Objective instability of the knee showed a significant difference when order group 1 was compared separately with order group 3 and with order group 4 (p = 0.004). Surgical time differed between surgeons (p < 0.001), and the shape of the learning curve showed great individual variability.

Conclusion: In the first 10 to 20 ACLR operations, the surgical time was longer and the complication rate higher, but thereafter both started to settle down. We recommend that first 10-20 ACLR operations should be supervised.