Purpose: We studied the construct validity of the LapMentor, a virtual reality laparoscopic surgical simulator, and the correlation between the data collected on the LapMentor and the results of video assessment of real laparoscopic surgeries.
Materials and methods: Ninety-two urologists were tested on basic skill tasks No. 3 (SK3) to No. 8 (SK8) on the LapMentor. They were divided into three groups: Group A (n=25) had no experience with laparoscopic surgeries as a chief surgeon; group B (n=33) had <35 experiences; and group C (n=34) had ≥35 experiences. Group scores on the accuracy, efficacy, and time of the tasks were compared. Forty physicians with ≥20 experiences supplied unedited videotapes showing a laparoscopic nephrectomy or an adrenalectomy in its entirety, and the videos were assessed in a blinded fashion by expert referees. Correlations between the videotape score (VS) and the performances on the LapMentor were analyzed.
Results: Group C showed significantly better outcomes than group A in the accuracy (SK5) (P=0.013), efficacy (SK8) (P=0.014), or speed (SKs 3 and 8) (P=0.009 and P=0.002, respectively) of the performances of LapMentor. Group B showed significantly better outcomes than group A in the speed and efficacy of the performances in SK8 (P=0.011 and P=0.029, respectively). Analyses of motion analysis data of LapMentor demonstrated that smooth and ideal movement of instruments is more important than speed of the movement of instruments to achieve accurate performances in each task. Multiple linear regression analysis indicated that the average score of the accuracy in SK4, 5, and 8 had significant positive correlation with VS (P=0.01).
Conclusions: This study demonstrated the construct and predictive validity of the LapMentor basic skill tasks, supporting their possible usefulness for the preclinical evaluation of laparoscopic skills.