PubMed abstract / permitted excerpt

OBJECTIVE: To investigate effects of repeated training with anatomically varied versus identical 3D-printed models on skill acquisition, cognitive load, and skill retention in novice mastoidectomy trainees. STUDY DESIGN: A prospective, randomized, controlled, educational trial conducted April to September 2024 including a 3-month retention test. SETTING: Simulation center at a single academic teaching hospital. METHODS: Twenty-four novice medical students were randomized 1:1 into 2 groups: The Identical Models (IM) group practiced repeated procedures on identical 3D-printed temporal bone models (control) and the Variable Model (VM) group practiced on a varied selection of models (intervention). Trainees completed 10 anatomical mastoidectomy procedures with posterior tympanotomy followed by 2 retention test procedures after 3-months of nonpractice. The primary outcome was performance assessed with 25-point modified Welling Scale. The secondary outcome was cognitive load estimated by relative reaction time measured on a secondary task. RESULTS: At end-of-training, the VM group scored 20.6 points (95% CI [19.2-22.0]), outperforming the IM group who scored an average of 18.2 points (95% CI [16.7-19.6], P = .02). At 3‑month retention testing, the VM group's scores decayed by a mean of 5.6 points on the repeated model (95% CI [-7.2 to-4.0], P < .001) and of 3.7 points when introduced to a new model (95% CI [-5.2 to-2.1], P < .001). The IM group showed no decay thereby demonstrating better skill retention which aligned with reduced relative reaction time (cognitive load) during training. CONCLUSIONS: Educators can optimize mastoidectomy skills acquisition and retention by instructional design that prioritize repeated practice on anatomically identical models until basic skills have been acquired.