Construct validity of a virtual reality simulator for surgical training in knee arthroscopy

Abstract

Objective

Surgical techniques are learned gradually throughout an orthopedic residency. Training on real patients carries drawbacks such as limited access and elevated risk. Alternatively, surgical simulation allows residents to practice in a safe environment with greater access to standardized surgical tasks. Virtual reality simulators display images inside an artificial joint, often providing real-time haptic feedback to allow for realistic interaction. The objective of this study was to evaluate the construct validity of a virtual reality simulator for knee arthroscopy by analyzing the capacity of system parameters to distinguish between expert and novice surgeons.

Design

This comparative cross-sectional study contrasts the automated performance reports for novice and expert orthopedic surgeons after executing surgical tasks on the ARTHRO Mentor virtual reality simulator.

Setting

Surgical simulation center at the University of Chile Clinical Hospital, Santiago, Chile.

Participants

The novice group consisted of 20 second-year orthopedic and traumatology residents at the University of Chile School of Medicine. The expert group consisted of 10 experienced arthroscopic surgeons. All participants carried out standardized tasks in the knee arthroscopy virtual reality simulator. The median performance scores of the two groups were compared, and multivariate logistic regression was performed to assess the capacity of the system to discriminate between the two groups.

Results

Median performance on the vast majority of surgical tasks was superior for the expert group. The expert group had performance values equal to or higher than the novice group on 43 of the 44 variables recorded for the basic tasks and 74 of the 75 advanced task variables. The multivariate logistic regression analysis discriminated expert from novice users with 100% accuracy.

Conclusion

The virtual reality simulator for knee arthroscopy showed good construct validity, with performance metrics accurately discriminating between expert and novice users.