LAUSR

January 28, 2020 237 James A. de Lemos, MD Brahmajee K. Nallamothu, MD T his week’s issue of Circulation includes 3 papers that we expect will generate considerable controversy. These studies use observational study designs to compare the effectiveness of different, competing cardiac procedural interventions. In 1 study, Amin et al compare contemporary usage patterns and clinical outcomes of the Impella intravascular microaxial blood pump versus the intra-aortic balloon pump (IABP) in patients undergoing percutaneous coronary intervention with mechanical circulatory support.1 The other 2 studies, by Van Belle et al and Deharo et al, compare clinical outcomes of balloon-expandable versus self-expanding bioprosthetic valves for use in transcatheter aortic valve replacement (TAVR).2,3 Despite calls for more “real-world evidence” from observational comparative effectiveness (CE) studies, including from regulatory agencies, Circulation publishes relatively few papers using these methods. Most of our editors believe that even the very best methodologies for observational CE research are insufficient to fully address intrinsic limitations related to bias and confounding. The most robust observational CE methods, including instrumental variable analysis, propensity matching, and multivariable adjustment, aim to recapitulate the experimental conditions of randomized controlled trials (RCTs). However, this is not possible to achieve with total confidence, and observational CE research studies may reach conclusions that fundamentally differ from RCTs later conducted on the same topic. This is critical because, without supportive evidence from RCTs, we can never truly know when the observational study gets it right, or when it gets it wrong. In the situation where observational CE studies reach different conclusions from previously conducted large RCTs on the same topic, the RCT result is invariably more valid and reliable, unless there were major methodological problems with the RCT. Given our inherent skepticism of conclusions drawn from observational CE research, why are we publishing these papers? What criteria do the editors use to assess this type of research in Circulation? Despite the important limitations outlined earlier, there is a valuable role for observational CE research in Circulation, but we believe the scope is fairly narrow. We approach these types of analyses with the following considerations in mind: 1. The CE question must be scientifically important and clinically relevant. 2. The study question should be unanswered. If the hypothesis has already been adequately tested using a superior RCT study design, we would need a compelling justification for the observational CE study, which may include the following: a. Major changes in natural history or concomitant therapies for the disease being studied have occurred, or the observational CE research is examining the study question in different settings to examine real-world use and effectiveness.4 © 2020 American Heart Association, Inc. EDITOR’S PAGE