Eye Movement Modeling Examples in Robotic Surgery Training—A Randomized Controlled Study

摘要

• In a randomized controlled trial of 137 medical students, viewing expert eye movements (EMME) during robotic surgery training did not improve students' ability to interpret anatomical landmarks in laparoscopic images compared to a control group. • Students exposed to EMME spent significantly more time engaging with the training material (971 seconds vs 881 seconds), though this increased engagement did not translate to better performance. • The findings challenge previous assumptions about EMME's effectiveness in medical education, suggesting that additional visual cues may compete with rather than complement the intrinsic visual demands of laparoscopic imagery, particularly for advanced learners. In robotic-assisted laparoscopic surgery, the perception and interpretation of the visual scene poses a huge challenge for medical trainees. However, their ability to assist effectively in surgical procedures depends heavily on mastering these visual skills. Research on multimedia learning suggests that seeing an expert's eye movements during task completion as a “worked-example” (i.e., eye movement modeling examples [EMME]) may guide learners’ attention to critical areas and improve learning outcomes. The effectiveness of EMMEs in robotic surgical training remains unclear. A randomized controlled 2-group study was conducted in which participants viewed an audio-guided robotic prostatectomy video. The EMME group viewed the surgical video with an overlay showing an expert urologist's eye movements captured through eye-tracking technology, while the control group viewed the video without the overlay. Learning outcomes were assessed through anatomical landmark identification tasks in 14 laparoscopic images (reliability coefficient α = 0.684). Training engagement was measured through time spent with the learning material. The study was conducted within the medical education program at Charité Berlin in an affiliated secondary care geriatric hospital during 2024. The study included 137 fourth-year medical students randomized into 2 groups: EMME group ( n = 67; mean age 26 ± 4 years; 62.7% female) and control group ( n = 70; mean age 26 ± 5 years; 62.9% female). The EMME group engaged with the material significantly longer ( M = 971 s ± 241 s) compared to the control ( M = 881 s ± 217 s), t( 135) = 2.288, p = 0.012, d = 0.391, 95% CI [12.137; 167.066]. However, there was no significant difference in the ability to interpret anatomical landmarks in laparoscopic images between the EMME ( M = 54.9% ± 19.4%) and the control group ( M = 57.1% ± 18.7%), t( 135) = 0.688, p = 0.493, 95% CI [−0.042; 0.087]. Contrary to literature, these findings suggest that although EMME may enhance attention to educational content, the increased engagement does not automatically translate into superior visual performances in laparoscopic surgery. We discuss these results in the broader context of multimedia learning.