We strive to employ rigorous scientific and novel analytical methods in our work to produce the highest quality cutting edge research. We are committed to creating work that has real-world validity, is centered in equity, and serves to enhance the well-being of every community.
Research
Use of Prediction Algorithms to Optimize Care
We develop prediction algorithms to identify individuals at risk for cardiovascular diseases and to guide clinical care. In particular, we have developed an algorithm to identify individuals with need for ultrasound enhancing agents on echocardiography that has since been externally validated. Additionally, we developed and externally validated an algorithm to predict long-term progression of mitral and tricuspid regurgitation. We have also evaluated the independent risk of frailty in individuals with severe aortic valve stenosis and the relationship of age-related basal septal hypertrophy on outcomes after transcatheter aortic valve replacement.
Impact of Policies and Patterns of Care on Health Outcomes
We are interested in the use of claims and other nationally representative data sources to evaluate patterns of care delivery, particularly as it relates to the impact of a policy or initiative and any untoward effects this may have on individual health outcomes. Specifically, we evaluated how specific wording in statements on public reporting of percutaneous coronary intervention (PCI) may have influenced use of PCI in those with cardiac arrest. Additionally, we evaluated rates and costs of non-Medicare short-term readmissions. Moreover, we identified age, racial, and sex-related disparities in the receipt of echocardiography and whether hospital variation in mechanical circulatory use was associated with improved patient outcomes.
Validity of Novel Data Sources for Real-World Evidence Generation
We evaluate the use of real-world data generated through healthcare operations such as administrative claims and echocardiographic report data to validly ascertain outcomes and diagnoses of interest. In particular, we have linked trial data from the CoreValve Pivotal studies to Medicare claims to identify the optimal set of claims to identify cerebrovascular events, non-stroke events, and have assessed if claims could validly be used to replace events adjudication in pragmatic trials.