OBJECTIVE: Optimal temporal surgical management of significant carotid stenosis and coronary artery disease remains unknown. Carotid endarterectomy (CEA) and coronary artery bypass (CABG) are performed concurrently (CCAB) or in a staged (CEA-CABG or CABG-CEA) approach. Using the Vascular Quality Initiative-Vascular Implant Surveillance and Interventional Outcomes Coordinated Registry Network-Medicare-linked dataset, this study compared operative and long-term outcomes after CCAB and staged approaches. METHODS: The Vascular Quality Initiative-Vascular Implant Surveillance and Interventional Outcomes Coordinated Registry Network dataset was used to identify CEAs from 2011 to 2018 with combined CABG or CABG within 45 days preceding or after CEA. Patients were stratified based on concurrent or staged approach. Primary outcomes were stroke, myocardial infarction (MI), all-cause mortality, stroke and death as composite (SD) and all as composite within 30 days from the last procedure as well as in the long term. Univariate analysis and risk-adjusted analysis using inverse propensity weighting were performed. Kaplan-Meier curves of stroke, MI, and death were created and compared. RESULTS: There were 1058 patients included: 643 CCAB and 415 staged (309 CEA-CABG and 106 CABG-CEA). Compared with staged patients, those undergoing CCAB had a higher preoperative rate of congestive heart failure (24.8% vs 18.4%; P = .01) and decreased renal function (14.9% vs 8.5%; P < .01), as well as fewer prior neurological events (23.5% vs 31.4%; P < .01). Patients undergoing CCAB had similar weighted rate of 30-day stroke (4.6% vs 4.1%; P = .72), death (7.0% vs 5.0%; P = .32), and composite outcomes (stroke and death, 9.8% vs 8.5%; P = .56; stroke, death, and MI, 14.7% vs 17.4%; P = .31), but a lower weighted rate of MI (5.5% vs 11.5%; P < .01) vs the staged cohort. Long-term adjusted risks of stroke (hazard ratio [HR], 0.85; 95% confidence interval [CI], 0.54-1.36; P = .51) and mortality (HR, 1.02; 95% CI, 0.76-1.36; P=.91) were similar between groups, but higher risk of MI long-term was seen in those staged (HR, 1.49; 95% CI, 1.07-2.08; P = .02). CONCLUSIONS: In patients undergoing CCAB or staged open revascularization for carotid stenosis and coronary artery disease, the staged approach had an increased risk of postoperative cardiac event, but the short- and long-term rates of stroke and mortality seem to be comparable. Adverse cardiovascular event risk is high between operations when staged and should be a consideration when selecting an approach. Although factors leading to staged sequencing performance need further clarity, CCAB seems to be safe and should be considered an equally reasonable option.

BACKGROUND: Validating new algorithms, such as methods to disentangle intrinsic treatment risk from risk associated with experiential learning of novel treatments, often requires knowing the ground truth for data characteristics under investigation. Since the ground truth is inaccessible in real world data, simulation studies using synthetic datasets that mimic complex clinical environments are essential. We describe and evaluate a generalizable framework for injecting hierarchical learning effects within a robust data generation process that incorporates the magnitude of intrinsic risk and accounts for known critical elements in clinical data relationships. METHODS: We present a multi-step data generating process with customizable options and flexible modules to support a variety of simulation requirements. Synthetic patients with nonlinear and correlated features are assigned to provider and institution case series. The probability of treatment and outcome assignment are associated with patient features based on user definitions. Risk due to experiential learning by providers and/or institutions when novel treatments are introduced is injected at various speeds and magnitudes. To further reflect real-world complexity, users can request missing values and omitted variables. We illustrate an implementation of our method in a case study using MIMIC-III data for reference patient feature distributions. RESULTS: Realized data characteristics in the simulated data reflected specified values. Apparent deviations in treatment effects and feature distributions, though not statistically significant, were most common in small datasets (n < 3000) and attributable to random noise and variability in estimating realized values in small samples. When learning effects were specified, synthetic datasets exhibited changes in the probability of an adverse outcomes as cases accrued for the treatment group impacted by learning and stable probabilities as cases accrued for the treatment group not affected by learning. CONCLUSIONS: Our framework extends clinical data simulation techniques beyond generation of patient features to incorporate hierarchical learning effects. This enables the complex simulation studies required to develop and rigorously test algorithms developed to disentangle treatment safety signals from the effects of experiential learning. By supporting such efforts, this work can help identify training opportunities, avoid unwarranted restriction of access to medical advances, and hasten treatment improvements.