Documenting a learning curve and test-retest reliability of two tasks on a virtual reality training simulator in laparoscopic surgery.

Background: Virtual reality simulators are a component of the armamentarium for training surgical residents. No one knows exactly how to incorporate virtual reality simulators into a curriculum. The purpose of this study was to document and show the learning curve and test-retest reliability of 2 tasks on a virtual reality-training simulator (LapSim; Surgical Science, Göteborg, Sweden) in laparoscopic surgery.

Methods: Twenty-nine medical students participated in 8 iterations of 7 virtual reality tasks ("camera navigation" (CN), "instrument navigation," "coordination," "grasping," "lifting and grasping" (LG), "cutting," and "clip applying") Learning curves for each outcome variable of the CN and LG tasks were generated. Using ANOVA, we evaluated the differences between each score from attempt number 7 to attempt number 8 to document test-retest reliability.

Results: A plateau in the learning curve occurred within 8 sessions for CN misses, CN tissue damage, CN maximum damage, and LG maximum damage. Over the course of 8 sessions, a plateau in the learning curve was nearly reached for CN time, CN drift, CN path, CN angular path, and LG left and right path. The following variables had a downward trend to the mean learning curve over 8 sessions, but they did not reach a plateau: LG time, LG left and right miss, LG left and right angular path, and LG tissue damage.

Conclusion: Using the LapSim virtual reality simulator, we documented a learning curve and test-retest reliability for each outcome variable for CN and LG for rank novices. The modeling of the general learning curve is useful in designing training program. These results may be important in developing standards for technical evaluation in a surgical training curriculum.