Abstract
STUDY DESIGN: A systematic review and meta-analysis of individual participant and aggregated data. OBJECTIVES: To define the learning curves of endoscopic discectomies using unified statistical methodologies. METHODS: Searches returned 913 records, with 118 full-text articles screened. Studies of endoscopic lumbar spine surgery reporting outcomes by case order were included. Mixed-effects nonlinear, logistic, and beta meta-regressions prdwere conducted to define the learning curves. RESULTS: 13 studies involving 864 patients among 15 surgeons were included in total. For transforaminal endoscopic discectomy, the estimated operating time for the first case was 95 min [CI: 87-104], and the estimated plateau was 66 minutes [CI: 51-81]. An estimated 21 cases [CI: 18-25] were required to overcome 80% of this deficit, but near-plateau performance was expected only after 59 cases [CI: 51-70]. The estimated risk of surgical complications on the first case was 25% [CI: 11%-46%], with an 80% reduction in relative risk requiring an estimated 41 cases. The expected postoperative VAS leg pain score after the first case was 2.7 [CI: 1.8-3.8], with an 80% improvement requiring an estimated 96 cases. Similar numbers were required to overcome the learning curves in interlaminar and biportal endoscopic discectomies. CONCLUSIONS: Approximately 60 cases are required to achieve proficiency in endoscopic lumbar spine surgery, although the greatest part of the learning curve can be overcome with 20 cases. This should be considered when designing implementation programmes for surgeons and service providers that wish to incorporate endoscopic spinal surgery into their practice.