Uses of equipoise in discussions of the ethics of randomized controlled trials of COVID-19 therapies

BACKGROUND: Early in the COVID-19 pandemic, the urgent need to discover effective therapies for COVID-19 prompted questions about the ethical problem of randomization along with its widely accepted solution: equipoise. In this scoping review, uses of equipoise in discussions of randomized controlled trials (RCT) of COVID-19 therapies are evaluated to answer three questions. First, how has equipoise been applied to COVID-19 research? Second, has equipoise been employed accurately? And third, do concerns about equipoise pose a barrier to the ethical conduct of COVID-19 RCTs? METHODS: Google Scholar and Pubmed were searched for articles containing substantial discussion about equipoise and COVID-19 RCTs. 347 article titles were screened, 91 full text articles were assessed, and 48 articles were included. Uses of equipoise were analyzed and abstracted into seven categories. RESULTS AND DISCUSSION: Approximately two-thirds of articles (33/48 articles) used equipoise in a way that is consistent with the concept. They invoked equipoise to support (1) RCTs of specific therapies, (2) RCTs in general, and (3) the early termination of RCTs after achieving the primary outcome. Approximately one-third of articles (15/48 articles) used equipoise in a manner that is inconsistent with the concept. These articles argued that physician preference, widespread use of unproven therapies, patient preference, or expectation of therapeutic benefit may undermine equipoise and render RCTs unethical. In each case, the purported ethical problem can be resolved by correcting the use of equipoise. CONCLUSIONS: Our findings highlight the continued relevance of equipoise as it supports the conduct of well-conceived RCTs and provides moral guidance to physicians and researchers as they search for effective therapies for COVID-19. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12910-021-00712-5.