Construct validity of five phacoemulsification modules on the EyeSi surgical simulator: a cross-sectional study across three levels of surgical expertise.

PURPOSE: Virtual reality simulators are increasingly integrated into cataract surgery training to facilitate skill acquisition outside the operating room. However, their educational value depends on the validity of simulated tasks. This study aimed to evaluate the construct validity of five procedural phacoemulsification modules on the EyeSi simulator (version 3.10.67) across three levels of surgical expertise.

METHODS

This cross-sectional comparative study was conducted at a tertiary referral hospital and a university simulation center in Morocco. Twenty-five participants were stratified into novices (n = 17), advanced residents (n = 4), and expert surgeons (n = 4). Participants completed five EyeSi phacoemulsification modules, and performance metrics were compared across groups.

RESULTS

Overall simulator performance improved with increasing surgical experience. Total scores differed significantly between groups (p = 0.005), with median values of 311, 365, and 473 for novices, advanced residents, and experts, respectively. Total completion time also differed significantly (p = 0.047). At the module level, construct validity was demonstrated for capsulorhexis (p = 0.030), divide and conquer (p = 0.009), and intraocular lens (IOL) insertion (p = 0.002), whereas hydrodissection and irrigation/aspiration did not reach statistical significance (p = 0.084 and p = 0.053, respectively).

CONCLUSIONS

In a recent EyeSi software version, construct validity was confirmed for three of the five evaluated procedural modules across three levels of surgical expertise. Complementary analyses further supported an association between simulator performance and surgical experience across all modules. These findings reinforce the educational value of the EyeSi simulator while highlighting the need for the continuous evaluation and refinement of individual procedural modules and their scoring algorithms.