A Novel Simulation Model and Training Program for Minimally Invasive Surgery of Hallux Valgus.

Abstract

Minimally invasive surgery (MIS) for hallux valgus (HV) has gained popularity. However, adopting this technique faces the challenges of a pronounced learning curve. This study aimed to address these challenges by developing and validating an innovative simulation model and training program, targeting enhanced proficiency in HV MIS. A training program and a high-fidelity simulation model for HV MIS were designed based on experts' recommendations. Four foot and ankle surgeons without experience in MIS formed the novice group and took the program that encompassed six-session instructional lessons, hands-on practice on simulated models, and immediate feedback. The program concluded with a cadaveric surgery. Four foot and ankle experienced MIS surgeons formed the expert group and underwent the same procedure with one simulated model. Participants underwent blind assessment, including Objective Structured Assessment of Technical Skills (OSATS), surgical time, and radiograph usage. Expert evaluation of the simulation model indicated high satisfaction with anatomical representation, handling properties, and utility as a training tool. The expert group consistently outperformed novices at the initial assessment across all outcomes, demonstrating OSATS scores of 24 points (range, 23 to 25) versus 15.5 (range, 12 to 17), median surgical time of 22.75 minutes (range, 12 to 27) versus 48.75 minutes (range, 38 to 60), and median radiograph usage of 70 (range, 53 to 102) versus 232.5 (range, 112 to 280). Novices exhibited a significant improvement in OSATS scores from the fifth session onward (P = 0.01), reaching the desired performance of 20 points. Performance at the final training with the simulated model did not differ from cadaveric surgery outcomes for all parameters. This study validated a simulation model and training program, allowing nonexperienced HV MIS foot and ankle surgeons to enhance their surgical proficiency and effectively complete a substantial portion of the learning curve at the fifth session, and this performance was successfully transferred to a cadaver model. III.