P. Sirimanna1, R. Aggarwal2, M. A. Gladman1 1University Of Sydney,Sydney Medical School – Concord,Sydney, NSW, Australia 2McGill University,Department of Surgery,Montreal, QC, Canada
Introduction: Laparoscopic appendectomy (LA) is the commonest emergency surgical procedure and is associated with significant variation in practice and outcomes. Further, this index procedure has a significant learning curve and is often a surgical trainee’s first experience of laparoscopy. Virtual reality (VR) simulation allows supplementary technical skill acquisition outside of the operating room (OR) with demonstrable translational benefits to actual OT performance. Surprisingly, training in LA has received little attention. Thus, we aimed to validate a VR simulator model of LA as a training and assessment tool, and obtain benchmarks of proficient skill in order to develop a proficiency-based VR simulation technical skills training curriculum for LA.
Methods: 10 experienced (performed >50 LAs), 8 intermediate (10-30 LAs) and 20 inexperienced (<10 LAs) operators were recruited. Using a high-fidelity VR laparoscopic surgical simulator (LAP MentorTM), experienced and intermediate operators performed 2 repetitions of 5 guided procedural modules that navigate users through the component steps of LA using a variety of surgical techniques, as well as an unguided module that allow a full LA to be practiced without direction. Inexperienced operators performed 10 repetitions of these modules to facilitate learning curve analysis. Performance was assessed using simulator-derived metrics of economy of motion, No of movements, path length, idle and task time. Construct validity was determined by comparing these metrics between the 3 groups for each module. Validated modules were used for curriculum construction, as were proficiency benchmarks for each metric (set as the median values obtained in the ‘experienced’ group) to provide performance goals.
Results: Guided modules demonstrated construct validity as evidenced by statistically significant differences between the 3 groups in terms of No of movements, path length, idle and task time (p<0.05), with statistically learning curves that plateaued between the 6th and 9th sessions (p<0.01). 2 of the 5 guided modules also exhibited construct validity for economy of motion. The unguided full LA module demonstrated construct validity for economy of motion (5.00 vs 7.17 vs 7.84, P<0.01), No of movements (1101 vs 690.5 vs 532, P<0.01), path length (1797.08cm vs 1573.51cm vs 1315.09cm, P<0.01), idle time (325.43s vs 160.44s vs 118.45s, P<0.01) and task time (864.49s vs 477.2s vs 352.12s, P<0.01), with statistically learning curves that failed to plateau. Examples of proficiency criteria used for curriculum construction include: task time (352s), path length (1315cm), and No of movements (532) for the unguided module.
Conclusions: A VR simulator model of LA was demonstrated to be a valid training and assessment tool. Consequently, the first evidence-based technical skills training curriculum for LA was constructed that facilitates acquisition of proficiency status by trainees.