Table 1 Considerations for the choice of a restricted randomization procedure

Mitigate potential for selection bias

A procedure should have high degree of randomness.

Mitigate potential for chronological bias.

A procedure should balance treatment assignments over time.

Valid and efficient treatment comparison

A procedure should have established statistical properties, provide strong control of false positive rate and yield unbiased, low variance estimates of the treatment difference.

A procedure should preserve the unconditional allocation ratio (e.g. 1:1) at every allocation step and achieve approximately or exactly the target sample sizes per group.

Ease of implementation

Validated statistical software for implementing a randomization procedure must be in place.