Abstract
Seamless phase 2/3 study designs provide a framework for a more efficient trial. During the COVID-19 pandemic, such study designs were considered particularly appealing as there was an urgent global need to rapidly identify effective therapeutics. However, limited in vivo safety and efficacy data was available early in the pandemic to inform decisions. As part of the ACTIV-2 study, we implemented a phase 2/3 platform trial to evaluate multiple candidate treatments for non-hospitalized adults with COVID-19. In addition to an adequate safety profile, the decision to graduate an agent from phase 2 to phase 3 was based on showing treatment effects on clinical or laboratory markers. Decision criteria evolved over time as more data became available during the global pandemic. A seamless transition and approximately 20 % reduction in total sample size was achieved for one agent, amubarvimab plus romlusevimab. Using both simulation studies and actual results from graduation assessments of five ACTIV-2 candidate therapeutics, we provide a discussion of lessons learned from our implementation and recommendations for future seamless trials of interventions for emergent infections.