To summarize the performance and implementation status of various randomization methods, and to provide clinical researchers with a guide for choosing an appropriate randomization method and executing it in a standardized manner.

The review elucidated the conceptual and implementation integrity of randomization, and illuminated the developmental history and application of major randomization methods, as well as their performance in the two competitive dimensions of group allocation randomness and group sample balance.

Restricted randomization was frequently used to maintain an appropriate treatment balance in practice. Block randomization was the most widely used, but its allocation sequence lacked sufficient randomness and could be predicted, elevating the risk of selection bias, especially in open-label trials. Methods offering a superior trade-off between randomness and balance included MTI restricted randomization methods like the big stick design and maximal procedure, as well as combinations of multiple randomization methods such as the block urn design and sandwich mixed randomization. Many clinical researchers still had misconceptions about the randomization concept itself and often overlooked its critical methodological rigor and standardized implementation.

Researchers should meticulously choose a suitable randomization method and report implementation details in a standardized manner, allowing peer researchers to evaluate the internal validity and evidence strength of the study. Due to its high predictability, the widely used block randomization method should be substituted by alternative methods with better performance. Future randomization procedures can be explored and developed based on the MTI restrictive methods and the combinations of existing methods.