Refining Risk, Curating Care: A Data-Driven Approach to Carotid Revascularization

Abstract

In Part I, we establish a baseline by analyzing temporal trends and determinants of stroke risk in patients with medically treated asymptomatic carotid stenosis. By evaluating these trends in Chapter 2, we provide a reference point that puts our findings in the subsequent chapters into perspective. In Part II, we evaluate the limitations of relying on in-hospital data versus 30-day follow-up. Chapter 3 contextualizes how relying solely on in-hospital data can underestimate the perioperative rates of adverse events following carotid revascularization. By showing the proportion of perioperative adverse events that occur post-discharge, we provide context for subsequent chapters that rely on in-hospital outcomes. In Part III, we highlight the importance of symptomatic versus asymptomatic stratifications in clinical guidelines for determining treatment strategies and assessing patient risk. In Chapter 4, we investigate current guideline recommendations by stratifying patients who underwent tfCAS into cohorts based on their specific symptom status. Building on the insights from the previous chapter, Chapter 5 extends the analysis to TCAR, a novel minimally invasive intervention procedure that had not been previously analyzed in this context. By examining outcomes based on specific preprocedural symptoms, these chapters address a research gap and reinforce the need for refined patient stratification in both tfCAS and TCAR. In Part IV, we leverage registry data to provide an individualized, patient-level view on best practices. Chapter 6 examines highly specific patient cohorts by investigating the optimal timing of CEA after stroke, stratified by stroke severity using the modified Rankin Scale. By focusing on CEA, the most commonly used method of carotid revascularization, the study offers a granular analysis of how intervention timing affects outcomes for patients with varying degrees of disability. Chapter 7 expands the focus beyond CEA, comparing optimal timing across multiple revascularization modalities: CEA, tfCAS, and TCAR. Part V looks into the future by addressing the fundamentals. As administrative and registry datasets are increasingly integrated for quality improvement and used in research, data accuracy becomes even more important. Chapter 8 evaluates the potential underreporting of stroke in these datasets by introducing a new metric that utilizes readily available data, the ratio of strokes to deaths, to assess and enhance the validity of outcomes based on administrative and registry data.