关键词: 靶向药物, 全球同步研发项目, 检验效能, 富集设计, 全随机设计

Abstract:

AIM: To evaluate the statistical performance of all-randomized design and enrichment design in simultaneous global development program for targeted drugs, and to provide a framework for the clinical practice. METHODS: Monte Carlo techniques were applied to simulate clinical data of simultaneous global development design for targeted drugs, and the power of all-randomized design and enrichment design was compared with different parameter settings.RESULTS:When the efficacy of non-targeted ethnic (NTE) patients and targeted ethnic (TE) patients, target positive and negative patients were equal, the power of all-randomized design and enrichment design with allocation was the highest, and the power of the enrichment design with screening was lower. When the target negative population had lower efficacy than the positive population, the enrichment design with allocation had the highest power, the second was all-randomized design, and the enrichment design with screening was the lowest. The power rose with the ratio of target positive population increased. When the efficacy of TE patients was lower than that of NTE patients, the results of power were almost the same with those in above, and the power was slightly lower. CONCLUSION: When there is sufficient evidence of a better effect on targeted positive patients than negative patients, the enrichment design is recommended. When researchers can't make sure that target drug is effective for marker-negative patients or not or there is not a reliable target diagnosis method, all-randomized design is recommended. And weighted Z test can provide enough power to verify the hypothesis when the sample size of local clinical trials is small.

Key words: targeted drugs, simultaneous global development program, power, enrichment design, all-randomized design