Key result
A multi-modal deep learning model achieved performance comparable to expert human raters for assessing surgical knot-tying skills, demonstrating the best overall performance among the tested models.
Why the study?
Objective assessment of technical skill is a growing but resource-heavy element of surgical education, motivating the development of automated surgical skill assessment.
Does a multi-modal deep learning model achieve comparable performance to expert human raters in assessing surgical knot-tying skills in surgical trainees and faculty?
Does a multi-modal deep learning model achieve comparable performance to expert human raters in assessing surgical knot-tying skills in surgical trainees and faculty?
Multi-modal deep learning models can achieve state-of-the-art performance comparable to expert human raters in objective surgical skill assessment, potentially reducing the burden on training faculty.
No takes yet. Share an insight, caveat, or question.
May reduce faculty burden in surgical training; leaves open prospective validation before adoption.
Kasa et al. (2022) studied Surgical skill assessment (n=72). Multi-modal deep learning model vs. Expert human raters and unimodal deep learning models was evaluated on Skill assessment performance measured using mean squared error (MSE) and intraclass correlation coefficient (ICC) on the OSATS GRS. A multi-modal deep learning model achieved performance comparable to expert human raters for assessing surgical knot-tying skills, demonstrating the best overall performance among the tested models.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: