Evaluating the optimal training paradigm for transcarotid artery revascularization based on worldwide experience.

BACKGROUND: Transcarotid artery revascularization (TCAR) is a new hybrid approach to carotid artery revascularization. Proctored training on live cases is an effort-, time-, and resource-intensive approach to learning new procedures. We analyzed the worldwide experience with TCAR to develop objective performance metrics for the procedure and compared the effectiveness of training physicians using cadavers or synthetic models to that of traditional in-person training on live cases. METHODS: Physicians underwent one of three mandatory training programs: (1) in-person proctoring on live TCAR procedures, (2) supervised training on human cadavers, and (3) supervised training on synthetic models. The training details and information from all subsequent independently performed TCAR procedures were recorded. The composite clinical adverse events (ie, transient ischemic attack, stroke, myocardial infarction, death) and composite technical adverse events (ie, aborted procedure, conversion to surgery, bleeding, dissection, cranial nerve injury, or device failure, occurring within 24 hours were recorded). Four procedural proficiency measures were recorded: procedure time, flow-reversal time, fluoroscopy time, and contrast volume. We compared the adverse event rates between the procedures performed by physicians after undergoing the three training modes and tested whether the proficiency measures achieved during TCAR after training on cadavers and synthetic models were noninferior to proctored training. RESULTS: From March 3, 2009 to May 7, 2020, 1160 physicians had undergone proctored (19.1%), cadaver-based (27.4%), and synthetic model-based (53.5%) TCAR training and had subsequently performed 17,283 TCAR procedures. The proctored physicians had treated younger patients and more patients with asymptomatic carotid stenosis and had had more prior experience with transfemoral carotid stenting. The overall 24-hour composite clinical and technical adverse event rates, adjusted for age, sex, and symptomatic status, were 1.0% (95% confidence interval, 0.8%-1.3%) and 6.0% (95% confidence interval, 5.4%-6.6%), respectively, and did not differ significantly by training mode. The proficiency measures of cadaver-trained and synthetic model-trained physicians were not inferior to those for the proctored physicians. CONCLUSIONS: We have presented key objective proficiency metrics for performing TCAR and an analytic framework to assess adequate training for the procedure. Training on cadavers or synthetic models achieved clinical outcomes, technical outcomes, and proficiency measures for subsequently performed TCAR procedures similar to those achieved with training using traditional proctoring on live cases.