The pursuit of balance using stratified and dynamic randomization techniques: An overview

Many pharmaceutical industry trials seek to achieve a balanced allocation of treatments over prognostic factors, including center. Much of the literature has discussed the merits of this from the viewpoint of statistical efficiency. This motivation is discussed and evaluated for trials of varying sizes with particular reference to multicenter trials. It is concluded that statistical efficiency for the primary analysis variable will rarely be the main reason for seeking balance and other often more important reasons are outlined. Randomization and dynamic allocation techniques to achieve balance are described and evaluated; these are Zelen's method, stratified permuted blocks randomization, minimization, urn designs, and optimal allocation techniques. Practical considerations arising from experience with using the techniques are described. Criteria which the practitioner can use to judge which is the most appropriate technique for his particular trial are given; these are balance, predictability, power, implications for analysis, and implications for medication supply. System requirements for the implementation of these techniques are described.