Co-designing a Novel Ordinal Endpoint for an Adaptive Platform Trial, BANDICOOT, in Pediatric Hematopoietic Stem Cell Transplant

An adaptive platform trial (APT) offers the ability to incorporate several research questions in the same target population across multiple domains (interventions), with the ability to add new questions in a perpetual manner. An APT is particularly appealing for pediatric hematopoietic stem cell transplant (HCT); an area of high heterogeneity, limited trial availability, and high mortality. Ideally, all domains in an APT would have the same primary endpoint. Therefore, an ordinal endpoint with multiple categories that combines various clinical outcomes into a single outcome measure is particularly appealing for APTs. Unfortunately, there is no accepted ordinal endpoint for pediatric HCT trials. This article aims to describe the methodology used to co-design a novel primary ordinal endpoint for the pediatric HCT APT — BANDICOOT. BANDICOOT is a study that aims to build an adaptive novel platform design — improving the complications, cost-effectiveness, outcomes, and overall survival from hematopoietic stem cell transplantation.The results of this process identified two potential ordinal endpoints that could be used, one focusing on organ support and the other on a combination of organ support, viral reactivation, and immune reconstitution. We explored the data extraction required for these endpoints from electronic medical records that we will utilize to validate the endpoints and determine which will be used in the APT BANDICOOT. In an era in which APTs are becoming increasingly utilized to answer important questions in clinical care, this article describes a reproducible strategy for the design of high-quality and meaningful ordinal endpoints.