Block Randomisation

Block randomisation

Block randomisation is a restricted randomisation method used in randomized controlled trials (RCTs) to keep treatment arms balanced throughout enrolment. By allocating participants in small blocks (for example 4, 6, 8) that contain a pre‑specified mix of treatment assignments, the technique prevents substantial short‑term imbalances - an important safeguard when recruitment is staggered, sample sizes are modest, or early stopping/interim analysis is possible (NICE PMG10; GOV.UK).

How it works

• The allocation sequence is built from randomly permuted blocks.
• For a 1:1 two‑arm trial with block size 4, each block contains two A and two B assignments in a random order (possible within‑block sequences: AABB, ABAB, ABBA, BAAB, BABA, BBAA).
• Blocks are concatenated and used during recruitment so that after each block the numbers in each arm are balanced.
• Block size must be compatible with the allocation ratio (for example, use block sizes that are multiples of 3 for a 2:1 allocation).
• Blocking is often applied separately within strata (site, age band, disease severity) so that balance is maintained within key prognostic subgroups.

Fixed versus variable block sizes

• Fixed block sizes maintain the tightest short‑term balance during recruitment.
• Fixed block sizes become predictable if the size is known: after observing allocations within a block, the remaining assignment(s) may be inferred, creating a risk of selection bias.
• Variable (randomly varying)...