简单随机化、中心分层区组随机化和最小化法的均衡性比较

The Balance Comparison Among Simple Randomization, Stratified Blocked Randomization and Minimization

目的比较简单随机化、中心分层区组随机化和最小化法的均衡性。方法运用MonteCarlo方法分别进行三种随机化方法的模拟分组,然后比较三种随机化方法在有6个非处理因素时,组间总例数的均衡性及非处理因素组间分布的均衡性。结果中心分层区组随机化和最小化法可保证组间总例数的均衡;在保证非处理因素的组间分布均衡性上,最小化法效果最好,明显优于其他两种方法,中心分层区组随机化与简单随机化在保证非处理因素均衡性上效果接近。结论最小化法既可保证组间总例数的均衡也可保证非处理因素组间分布的均衡,因此在保证均衡性上是首选。

Objective To compare the balance of simple randomization, stratified blocked randomization and minimization. Methods Monte Carlo technique was employed to simulate the treatment allocation of simple randomization, stratified blocked randomization and minimization respectively, then the balance of treatment allocation in each group and the balance for every prognostic factor were compared. Results The simulation demonstrated that minimization provides the best performance to ensure balance in the number of patients between groups and prognostic factors. Balance in prognostic factors achieved with stratified blocked randomization was similar to that achieved with simple randomization. Conclusion Minimization offers the best balance in the number of patients and prognostic factors between groups.