Evaluating Learning Curves in Virtual Reality Cortical Mastoidectomy Training Across Expertise Levels.

Abstract

Hypothesis: Virtual reality (VR) simulation has been established as an effective method of supplementing traditional surgical training. Learning curves can analyze skill acquisition over time in VR settings. Although previous studies explored learning curves of mastoidectomy performances on single specimen, this study analyzed learning curves on anatomically different virtual temporal bones across three expertise levels (novice, intermediate, and expert).

Methods: Thirty participants were divided into three groups: 10 medical students (novice), 10 ear, nose and throat (ENT) registrars (intermediate), and 10 senior ear surgeons (expert). They performed mastoidectomy on eight anatomically different temporal bones on the University of Melbourne Temporal Bone Surgery Simulator. A blinded senior ENT surgeon assessed the final products of the dissections using the Melbourne Mastoidectomy Scale (MMS). Learning curves of MMS scores, number of strokes, average force, drilling time, and total time were compared between groups using Friedman tests. Within-group analyses were conducted with Wilcoxon signed-rank tests.

Results: All performance metrics showed significant differences across all groups. Only comparison between intermediate and expert groups for average force was not significant. Within-group analyses showed significant differences in expert group for total and drilling time, intermediate group for average force, and novice group for total time. Individual learning curves of intermediate group demonstrated varied learning behavior.

Conclusions: Trainee performance was not seen to reach the level of experts after eight repetitions on anatomically different specimens. Enhancing training with individualized feedback and increased repetitions may optimize skills acquisition.