Classification of Deaths in Cardiovascular Outcomes Trials: Known Unknowns and Unknown Unknowns

Abstract

In 2002, Donald Rumsfeld, then US Secretary of Defense, famously stated “There are known knowns. There are things we know that we know. There are known unknowns. That is to say, there are things that we now know we don’t know. But there are also unknown unknowns.”1 His comment captures the range of possibilities when information is uncertain; from knowable to unknowable. Also, in science, it is the unknown unknowns that present the greatest challenge. Randomized, controlled, blinded clinical trials remain the gold standard for evaluating efficacy and safety of new therapeutics. Their primary strength is the elimination of any potential bias in the comparison between the investigational treatment(s) and control(s). However, as for any experiment, the methods of conduct of the experiment affect the quality of the science and the ability to achieve an adequate test of the hypothesis as intended. In the case of clinical outcomes trials, this important principle plays out through trial operations, which inherently influence the science and the robustness of the experiment. In our experience, countless operational decisions in the conduct of a clinical trial affect the quality of the experiment. Considering 1 important domain, many clinical trials use clinical event committees (CECs) to provide consistent assessment (adjudication) of clinical outcomes based on trial specific definitions. In this issue of Circulation, Fanaroff and colleagues2 present new analyses elucidating variability in the rates of adjudicated, undetermined cause of death and provide an important example of this interaction between trial operations and the output of the experiment.