Bringing the Credibility Revolution to Observational Research in Cardiology

Abstract

In the Spring of 2010, Economists Joshua Angrist and Jrn-Steffen Pischke reflected upon the vast strides achieved in the quality of empirical work in their field. 1"Empirical economics has experienced a credibility revolution," they wrote, "with a consequent increase in policy relevance and scientific impact."This had occurred in large part due to stronger research design in observational studies, with advancement of methodological strategies better equipped to address threats to the validity of causal inference.More than a decade later, Angrist, along with Guido Imbens and David Card, were awarded the Nobel Prize in Economics for their development and use of natural experiments to understand causal effects of policies.Medicine, and cardiology in particular, can rightly be credited for the revolutionizing the use of science's most powerful empirical tool, the randomized controlled trial (RCT).However, despite their strengths, RCTs are unlikely to be conducted for the vast majority of decisions that clinicians make each day.Thus, the proliferation of data from electronic health records, registries, surveys, administrative claims, mobile technologies and wearable presents an unprecedented opportunity to better understand the consequences of our medical decisions.Despite this, however, the use of observational data to consistently produce plausible assessments of treatment benefit or harm for cardiovascular therapies, devices and policies has faltered.Fueled by the ubiquity of data, an explosion of medical journals, and the unchecked incentive to publish, cardiovascular observational research has descended more deeply into a credibility crisis. The Crux of the ProblemAngrist and Pischke argued that careful attention to research design drove the credibility revolution in economics, with specific attention to one of the key challenges for observational studies -what economists refer to as omitted variable bias.Omitted variable bias is synonymous with what is referred to in the medical research lexicon as confounding by indication or treatment selection bias.Medicine's choice of terminology is meaningfulresidual confounding in a clinical comparison is often a consequence of the intentionality of

Topics & Concepts