In early-phase clinical trials researchers want to find the largest dose of a new drug that limits the risk of side-effects to an acceptable level, to use as an upper limit on dosing in subsequent trials. In a paper published by members of the NIHR Statistics Group, an explanation is given for a type of modern trial design that finds this dose: Continual Reassessment Method (CRM).
CRM re-analyses all the data after each patient and chooses the dose for the next patient that is expected to provide some benefit whilst controlling the risk of side-effects. Old-fashioned designs often had a fixed sequence of doses and only analysed at the end of the study. The dose chosen to take forwards for future research is more often correct and more patients are treated at, or close to, the chosen dose if the CRM design is used. This prevents the failure of promising new drugs due to the incorrect dose being used, means more is known about the side effects before more patients take the drug and leads to new drugs becoming available quicker for clinical use.
The paper offers a practical ‘how to’ guide for trialists and statisticians using the CRM design for the first time. It presents the planning decisions that need to be made in advance and gives guidance on operational aspects.
The paper was the result of a sequence of workshops organised by the NIHR Statistics Group that brought together several leading statisticians from academia and industry. They were aware of the inefficiencies of traditional dose-finding study designs and the lack of usage of modern methods, including CRM. Another paper from an overlapping group of authors investigates from a survey the barriers to adopting modern trial designs. This paper gives guidance on how to use the CRM design.