Abstract
In June–July 2022, the Randomization Working Group (https://randomization-working-group.rwth-aachen.de) conducted an online survey on the current practices in the application of randomization in clinical trials. Of 145 unique respondents, 137 (94.5%) identified themselves as statisticians. The majority of respondents were from academia and pharmaceutical companies. Permuted block randomization, with or without stratification, was the most frequently chosen method of randomization for RCTs. Interactive Web-Based Response Systems (IWRS) and “in-house” or a combination of “in-house and outsourced” randomization models were found to be most common in practice. Over 80% of respondents perceived some challenges to adoption of new randomization methods that may have more desirable properties. Over 80% of respondents identified opportunities for improving current practice, including education/training, development of standards/guidance on randomization, and adoption of validated software for generating randomization sequences. In summary, practitioners acknowledge the pivotal role of randomization in clinical trials. There are some perceived challenges to successful implementation of randomization, and there are opportunities for improving practice.
Supplementary Materials
The survey structure, including questions and response choices, with snapshots of the web pages, can be found in Supplemental Appendix 1 at the journal website.
Acknowledgments
The authors wish to thank the two anonymous referees and the associate editor for their review and valuable feedback that helped improve the manuscript. We would also like to acknowledge those who administered the distribution lists of various organizations that agreed to distribute the survey invitation to subscribers, allowing us to improve the uptake of the survey. Finally, we are grateful to all those who participated in our survey—without their responses, this work would not have beenpossible.
Disclosure Statement
The authors report that there are no competing interests to declare.