Objectives: To determine the length of the learning curve of femtosecond laser-assisted cataract surgery (FLACS) by a risk-adjusted cumulative sum method. Materials and Methods: This was a retrospective review of the first 288 FLACS performed by three surgeons over a 12-month period. The learning curves were analyzed separately for each surgeon and as pooled data via risk-adjusted cumulative sum methods. Change-point analysis was performed to estimate the length of the learning curve. Results: Rates of complications were as follows: anterior capsular tear (ACT) (3.4%), posterior capsular rupture (PCR) (0.7%), vitreous loss (0.3%), iris trauma (0.7%), Descemet's membrane tear (0.3%). There was a strong confidence level (96%) that there was a stable performance in terms of PCR after case 16 for the pooled data. Surgeon 1 did not have any cases of PCR, surgeons 2 and 3 exhibited change points at case 5 and 16, respectively (confidence levels = 99%, 98%). There was a strong confidence level (99.8%) that there was stable performance in terms of ACT after case 14 for the pooled data, however the individual surgeons demonstrated individual change points at case 1 (surgeon 1, confidence 99%), case 14 (surgeon 2, confidence 99%), and case 49 (surgeon 3, confidence 96%). Conclusions: The results suggest that there is an increased risk of ACT and PCR in FLACS within a surgeon's first 14 and 16 operations, respectively. Measures should be adopted to minimize harm to patients during this phase. However, overall rates of complications were safe.
Keywords: CUSUM analysis; Cataract; femtosecond laser assisted cataract surgery; learning curve; phacoemulsification.