针对多档次试验结果的情况,本文构造了一种罐子模型,并在此模型中建立了极限理论.理论和例子表明,这种临床试验设计,既符合人道精神,又保持较好的统计效率.因此该文有应用参考价值.
A kind of adaptive designs in clinical trials with multi-out come have been constructed using urn models some limit theorems have been shown in such models, some examples are given, which indicate that the better treatment will be allocated to patients with large chance, and the statistical etfficiency in these designs is still keeping better. So these adaptive designs are feasible.