Randomization is the optimal approach to ensuring that groups are prognostically balanced. The larger the sample size, the greater the likelihood that randomization will succeed in the objective of achieving prognostic balance. Optimal successful randomization includes an appropriate method of generating the allocation sequence and a procedure (ideally central randomization) to ensure concealment of randomization. When sample sizes are small-to-moderate, randomization may fail to do its job of ensuring prognostic balance and constrained randomization schemes - stratification and blocking, or minimization - may ameliorate the problem. However, especially with the potential increased predictability, investigators may consider alternatives to blocking, such as the emerging MP.
Authors
Guyatt GH; Alexander PE
Book title
Important Considerations for Clinical Trial Methodologies