Work

Evaluation of complex clinical trial designs in the context of statistical methodology and regulatory acceptance

Year: 2026
Type: dissertation
Abstract: Background and objectives The use of complex clinical trials (CCTs) has become increasingly popular in recent years, offering a new framework for drug development by evaluating multiple treatments or patient populations within a single overarching trial. CCTs include basket trials, which investigate a single treatment across multiple disease types; umbrella trials, which evaluate multiple treatments within a single indication; matrix trials, which evaluate multiple treatments within multiple indications; and platform trials, which provide flexibility in adding and removing arms during an ongoing trial. CCTs often use a common control group, which can improve efficiency but also introduce additional methodological and regulatory complexity. These developments have created a need for a deeper understanding of how such designs affect statistical operating characteristics, trial conduct and regulatory acceptability. The main objective of this dissertation is therefore to contribute to the understanding of regulatory and methodological aspects of CCTs, and to provide insights into the current regulatory thinking and methodological challenges associated with these trials. In particular, the dissertation focuses on the effect of common controls in CCTs on type I error and power, and the consequences of information leakage in platform trials with survival endpoints. In addition, the regulatory implications and assessment of aspects inherent in CCTs is summarised. By combining methodological and regulatory perspectives, the dissertation aims to contribute to a comprehensive understanding of CCTs. Methods This dissertation is based on four publications that address complementary dimensions of regulatory and statistical aspects of CCTs. To understand regulatory requirements for CCTs, available regulatory guidelines, including regulatory and stakeholder experience were summarised. The aim was to identify critical issues related to the planning, conduct and assessment of CCTs. To broaden this perspective, a systematic review of scientific advice procedures submitted to the European Medicines Agency (EMA) was conducted to determine recurring concerns and comments from regulators in the evaluation of CCTs. In a thorough simulation study, the use of a common control in CCTs and its impact on the type I error rate and power was evaluated. In particular, it was of interest to understand the type I error rate and power for platform trials where an arm joins the trial at a later stage. A second simulation study was conducted to examine the implications for the type I error and trial integrity due to issues of information leakage in platform trials with survival endpoints. To provide an estimation of the type I error inflation due to the information leakage, the maximum type I error rate was examined. Results The summary of regulatory guidance and scientific advices reveal that a number of aspects inherent in CCTs have not been fully addressed in currently available european guidelines but are of relevant concern mentioned by regulators. This includes for example the requirement for multiplicity adjustment in CCTs or the use of non-concurrent controls. The simulation results show that the use of common controls in platform trials increases the efficiency, particularly through improved power. At the same time, the common control inflates the type I error rate as multiple hypotheses are tested simultaneously in platform trials. However, in comparison to a trial with individual control arms, the probability of at least one false positive finding is smaller with the common control, whereas the probability of at least two or more false positive findings is larger. This is due to the correlation of test statistics when using a common control, and the correlation depends on the degree of overlap in control patients between arms. Further, the simulation results demonstrated that information leakage represents an issue to the trial integrity of platform trials with survival endpoints and common controls. When treatment arms on a platform trial are completed, the dissemination of results can be informative for arms still on the trial. Comparative efficacy results can be estimated based on the leaked data and can inflate the type I error rate when used for decision making. The maximum type I error depends on the time point of information leakage and the type of leaked information. Conclusion and discussion In conclusion, this dissertation provides a comprehensive overview of the regulatory and methodological aspects of CCTs, with a focus on platform trials. The four publications show that CCTs offer important opportunities for increasing efficiency in drug development, but that these advantages come at a cost of increased methodological and regulatory demands. The results highlight the need for careful consideration of the type I error rate and power when using common controls in CCTs and raise awareness of the potential issue of information leakage in platform trials. Overall, this dissertation contributes to the ongoing discussion on the use of CCTs in drug development and a better understanding of various statistical aspects of CCTs. This can aid regulators and methodologists in future scientific research and to develop new guidelines.
Language: English
Open Access status: green
OpenAlex ID: W7219853886
Cites: 61
Cited by:
Related to: 9