Validation of a Simulation Model for Robotic Sacrocolpopexy

Abstract

Study Objective To validate a robotic sacrocolpopexy (RSCP) simulation model including presacral dissection/mesh attachment, vaginal mesh attachment, and peritoneal closure. Design Simulation study. Setting Simulation lab. Patients or Participants Convenience sample of 9 Female Pelvic Reconstructive Surgery (FPMRS) surgeons and 17 current FPMRS fellows. Interventions FPMRS surgeons and current FPMRS fellows were videotaped using the RSCP model to complete a RSCP. Videotaped sessions were timed and scored using the validated Global Evaluative Assessment of Robotic Skills (GEARS) by 3 surgeon reviewers, all who were masked to subjects’ identity. Construct validity was measured by comparing performance on the model between experienced surgeon experts and trainees. Inter-rater reliability was determined by calculating interclass correlation coefficients for total GEARS scores. Face validity was assessed by a post-procedure questionnaire. Measurements and Main Results Experts included 9 FPMRS (2 male, 7 female) surgeons; trainees included 17 fellows (4 male, 13 female). Experts practiced at 7 different institutions; most (5/7) taught fellows. Trainees were from 7 institutions with various years of training: PGY5 (n= 6), PGY6 (n= 5), and PGY7 (n= 6). The experts’ performances were rated significantly higher for total GEARS scores and relevant domains of the GEARS scale. The GEARS autonomy parameter was scored as 5 for both groups, as all participants successfully completed the procedure without assistance. Between the three reviewers, inter-rater reliability was 0.99 for pair 1, 0.96 for pair 2, and 0.99 for pair 3, indicating high inter-rater reliability. Regarding face validity, all subjects “agreed/strongly agreed” that the model closely approximated live RSCP surgery and was useful for teaching and learning the procedure. Conclusion Experts using this RSCP model consistently had higher scores in each GEARS domain when compared with trainees, demonstrating construct validity. Inter-rater reliability was high. Face validity was established, as all participants felt the model replicated live surgery and would be useful for surgical training and evaluation.