Abstract Presentation Examining the Influence of Analytic Methodology on Clinical Outcomes
Jeffrey Moses, MD, presents the abstract “A Propensity-matched Analysis of Impella Support in Patients Undergoing High-Risk PCI from a Large-scale Claims Database: Influence of Analytic Methodology on Clinical Outcomes.” This study sought to understand the limitations of administrative databases and the potential to generate misleading or inconsistent conclusions. The study presents results from data using the same Premier Health Database (PHD) as used in a 2019 analysis by Amin et al. in which results conflicted with more robust clinical evidence from randomized controlled trials and multi-center registries specifically designed to answer clinical questions. In contrast to Amin et al., this study, using propensity matching as a methodological approach to control for confounders, demonstrated that the risk of adverse clinical outcomes in patients undergoing Impella® assisted PCI was not increased compared to IABP, results consistent with randomized controlled clinical data from the PROTECT II study as well as results from an analysis of national trends in cardiogenic shock from Stretch et al.
Dr. Moses dives into the data and analytic methods and discusses inherent limitations of using such databases to make clinical inferences with the intent to inform clinical decisions.
Dr. Moses is the John and Myrna Daniels Professor of Cardiology and director of interventional cardiac therapeutics at Columbia University Medical Center as well as the director of complex coronary interventions at St. Francis Hospital in Roslyn, NY.