Design and validation of a low-cost, high-fidelity model for robotic pyeloplasty simulation training

Abstract

Owing to restrictions in operative experiences, urology residents can no longer solely rely on 'hands-on' operative time to master their surgical skills by the end of residency. Simulation training could help residents master basic surgical skills and steps of a procedure to maximize time in the operative room. However, simulators can be expensive or tedious to set up, limiting the availability to residents and training programs. The authors sought to develop and validate an inexpensive, high-fidelity training model for robotic pyeloplasty. Pyeloplasty models were created using Dragon Skin® FX-Pro tissue-mimicking silicone cast over 3-dimensional molds. Urology faculty and trainees completed a demographic questionnaire. The participants viewed a brief instructional videoand then independently performed robotic dismembered pyeloplasty on the model. Acceptability and content validity were evaluated via post-task evaluation of the model. Construct validity was evaluated by comparing procedure completion time, the Global Evaluative Assessment of Robotic Skills (GEARS) score, blinded subjective physical evaluation of repair quality (1-10 scale), and flow rate between experts and novices. In total, 5 urology faculty, 6 fellows, and 14 residents participated. The median robotic console experience among faculty, fellows, and residents was 8 years (interquartile range [IQR] = 6-11), 3.5 years (IQR = 2-4 years), and 0 years (IQR = 0-0.5 years), respectively. The median procedure completion time was 29min (IQR = 26-40min), and the median flow rate was 1.11mL/s (IQR = 0-1.34mL/s). All faculty had flow rates >1.25mL/s and procedure times <30min compared with 2 of 6 fellows and none of the residents (P<0.001). All faculty, half of the fellows, and none of the residents achieved a GEARS score ≥20, with a median resident score of 12.5 (IQR = 8-13) (P<0.001). For repair quality, all faculty scored ≥9 (out of 10), all fellows scored ≥8, and the median score among residents was 6 (IQR = 2-6) (P<0.001). The material cost was $1.32/model, and the average production time was 0.12 person-hours/model. This low-cost pyeloplasty model exhibits acceptability and content validity. Construct validity is supported by significant correlation between participant expertise and simulator performance across multiple assessment domains. The model has excellent potential to be used as a training tool in urology and allows for repetitive practice of pyeloplasty skills before live cases.