Technical efficiency in hip arthroscopy plateaus after approximately 240 cases while complication rates continue to decline: A 15‐year single‐surgeon series of 1490 consecutive procedures
Andrea Mei, Leonardo Clausetti, Martial Metrop, Guillaume Girardot, Nicolas BoninAbstract
Purpose
To define the learning curve of a single surgeon performing hip arthroscopy using traction time as the primary metric of technical efficiency, and complication rate as a secondary endpoint, in a cohort large and long enough to model the inflection point with objective change‐point methods.
Methods
A total of 1490 consecutive hip arthroscopies performed by one surgeon between January 2007 and April 2022 were retrospectively analysed. Cases were ordered chronologically and assigned a cumulative case number. Traction time was modelled against case number with Pearson correlation and a continuous two‐phase (broken‐stick) segmented regression; the breakpoint was estimated by grid‐search minimisation of the residual sum of squares, with a 95% confidence interval from non‐parametric bootstrap (400 resamples). A cumulative sum (CUSUM) chart was constructed. Experience phases were compared with the Welch t test, Mann–Whitney U test, and the two‐proportion test.
Results
Traction time decreased with experience ( r = −0.34; p < 0.001). Segmented regression identified a breakpoint at case 242 (95% confidence interval [CI], 208–258; model R 2 = 0.35): a first‐phase decline of 1.6 min per 10 cases was followed by a plateau near 35 min; the CUSUM peak (case 224) corroborated this transition. Mean traction time fell from 54.8 to 35.6 min before versus after the breakpoint ( p < 0.001). Graded with the validated modified Clavien–Dindo classification for hip preservation surgery, the patient‐level complication rate decreased across consecutive phases from 14.7% (cases 1–150) to 3.1% (151–500) to 1.5% (501–1490), and continued to decline beyond the efficiency plateau; most complications were Grade I.
Conclusions
In one of the largest single‐surgeon series reported, technical efficiency (traction time) stabilised at approximately 242 cases, whereas safety continued to improve well beyond this point. In this single‐surgeon experience, the learning curve extended over several hundred cases, well beyond the 20–30 cases reported in early small series; whether these thresholds apply to other surgeons and training contexts requires confirmation.
Level of Evidence
Level IV, therapeutic case series.