Learning curve and proficiency metrics for transcarotid artery revascularization

Background: When introduced to a new procedure, physicians improve their performance and reduce their procedural adverse event rates rapidly during the initial cases and then improvement slows, signaling that proficiency has been achieved. Determining when they have acquired proficiency has important implications for procedural innovation, education, credentialing, and patient safety. We analyzed the worldwide experience with transcarotid artery revascularization (TCAR), a hybrid approach to carotid revascularization, to identify the (1) procedural performance measures associated with clinical and technical adverse events; (2) target levels of performance measures that minimize adverse event rates; and (3) number of TCAR cases needed to achieve the target levels for the performance measures. Methods: The patient, lesion, and physician characteristics were collected for each TCAR procedure performed by each physician worldwide in an international quality assurance database. Four procedural performance measures were recorded for each procedure: flow-reversal time, fluoroscopy time, contrast volume, and total skin-to-skin time. 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, device failure), occurring within 24 hours were also recorded. Correlations between each performance measure and the clinical and technical adverse event rates were computed. The inflection points in the performance measures were identified at which no further improvements occurred in the adverse event rates. Finally, the minimum number of TCAR cases required to achieve the target performance measure levels was computed. Results: A total of 18,240 procedures performed by 1273 physicians were analyzed. Of the 18,240 patients, 34.9% were women and 62.5% were asymptomatic. The flow-reversal time correlated with clinical adverse events adjusted for age, sex, and symptomatic status (R-2 = 0.91; P<.0001) and adjusted technical adverse events (R-2 = 0.86; P<.0001). The skin-to-skin time correlated with adjusted technical adverse events (R-2 = 0.92; P<.0001). A reduction in flow-reversal times to <13.1 minutes and the skin-to-skin time to <81 minutes did not translate into further improvements in the adverse event rates. A minimum of 26 TCAR cases was required to achieve the target flow-reversal time, and a minimum of 15 cases was required to achieve the target skin-to-skin time. Conclusions: The flow-reversal time and skin-to-skin time are appropriate performance measures for establishing the level of expertise of physicians as they acquire skills to perform TCAR. A target time of <= 13.1 minutes for flow-reversal and 81 minutes for skin-to-skin time minimized the adverse event rates. Familiarity with the steps involved in performing TCAR was achieved after >= 15 cases, and minimizing clinical adverse events occurred after >= 26 cases.