What are dynamic methods of randomisation?

Dynamic methods of randomisation create the allocation sequence at the time of randomisation. This is in contrast to static methods that determine the allocation sequence in advance of the trial starting and record the sequence in a randomisation list.

Dynamic methods have some practical advantages compared to randomisation lists. For instance, many trials stratify by site and dynamic methods allow new sites to be added freely at any time, whereas a randomisation list must be extended before new sites can be added. A dynamic list will never "run out" of allocations, whereas a randomisation list must anticipate the maximum number of allocations needed per stratum. In addition, dynamic methods such as minimisation or urn randomisation provide better balance and can incorporate more prognostic factors than stratified methods.

However, because the randomisation schedule is not produced in advance of the trial, and due to some theoretical concerns about the appropriate method of analysis of data from trials using minimisation, we recommend you check with the trial statistician before choosing to use dynamic methods.