Blocked Randomization

What It is

Patients to be enrolled in a trial are taken a few at a time, depending on the number of study groups – this is called a “block.” Within this small number (typically <10), all possible orderings of the study groups are spelled out. (e.g., for a block size of four patients in a trial comparing treatments A and B it would be AABB, BBAA, ABAB, BABA, ABBA, BAAB). Blocks are randomly selected from these orderings to build the randomization list until all patients in a study group have an assignment.

Why It’s Important

As above, we cannot count on simple randomization to produce study groups with equal numbers of patients in them. This reduces the statistical power of the trial. Block randomization guarantees that the intended percentage (e.g., 50% experimental and 50% control) of patients in each study group will occur. To avoid revealing a subset of the randomization list, usually blocks of different sizes are created and selected for building the randomization list.
Another benefit is that, if inclusion or exclusion criteria are changed during the trial, the study groups will be evenly balanced throughout the trial.