Journal of Clinical Research & Bioethics

Subject randomization in multi-center trials: Better and simple alternatives to permuted block and minimization

6^th International Conference on Advanced Clinical Research and Clinical Trials

September 10-11, 2018 | Zurich, Switzerland

Wenle Zhao

Medical University of South Carolina, USA

Scientific Tracks Abstracts: J Clin Res Bioeth

Abstract:

Statement of the Problem: Subject randomization is the foundation for randomized controlled trials being the gold standard for efficacy assessment in clinical researches. Permuted block randomization and minimization are the most commonly used methods for controlling imbalances in the size and baseline covariate distributions between treatment groups, at the cost of allocation randomness. The elevated allocation predictability of permuted block and minimization increases the risk of selection bias. Several new randomization designs with optimal statistical properties have been published, but gained little application, due to their complexity in implementation. Methodology & Theoretical Orientation: The block urn design drops the enforced block end balance in the permuted block design while maintaining the same control on the maximum tolerated imbalance. It significantly reduces the proportion of deterministic assignment. The mass-weighted urn design allows any desired unequal allocations to be accurately targeted. Both the block urn design and the mass-weighted urn design use the urn model so that the implementation is just as easy as the permuted block design. The minimal sufficient balance method changes the attitude toward covariate imbalance. Instead of applying a zero-tolerance rule, it uses complete random assignment unless covariate imbalance becomes serious. In that scenario, a biased coin assignment is used favoring reducing the imbalances. Findings: For two-arm equal allocation with a block size of 6, block urn design has 5.9% deterministic assignment, compared to 25% for the permuted block design. The mass-weighted urn design can accurately target allocations like 1:???2:???3, rather than using approximate allocations or large block sizes, such as 10:14:17. The minimal sufficient balance method prevent serious imbalances in more than 10 covariates while maintain a high level of allocation randomness. Conclusion & Significance: Replace inferior permuted block and minimization with better alternatives is feasible and will enhance the credibility and quality of clinical trials. Recent Publications 1. Zhao W (2014) A better alternative to stratified permuted block design for subject randomization in clinical trials. Stat Med. 33(30):5239-48. 2. Zhao W, Hill M D and Palesch Y (2015) Minimal sufficient balance ??? a new strategy to balance baseline covariates and preserve randomness of treatment allocation. Stat Methods Med Res. 24(6):989???1002. 3. Zhao W (2015) Mass weighted urn design ??? a new randomization algorithm for unequal allocations. Contemporary Clinical Trials 43:209-216. 4. Zhao W (2016) A better alternative to the inferior permuted block design is not necessarily complex. Statistics in Medicine 35(10):1736-8. 5. Zhao W and Ramakrishnan V (2016) Generalization of Wei???s urn design to unequal allocations in sequential clinical trials. Contemp Clin Trials Commun 2:75-79.

Biography :

Wenle Zhao is a Professor of Biostatistics in the Department of Public Health Sciences at the Medical University of South Carolina, USA. He obtained his PhD in Biostatistics at MUSC in 1999, and worked in the clinical trial field since then. He is the Associate Director of the Data Coordination Unit at MUSC, which is the National Statistical and Data Management Center for three NIH funded clinical trial networks. His expertise focused on subject randomization design and clinical trial information management software development. The new randomization designs he published in recent years have all been implemented in large multi-center trials, and are well received by investigators. He has been actively presenting his research in international conferences, including more 34 invited talks.

E-mail: zhaow@musc.edu