Department of Biostatistics and Medical Informatics, University of Wisconsin-Madison, WI, U.S.A.
Stat Med. 2012 Nov 10;31(25):2944-54. doi: 10.1002/sim.5405. Epub 2012 Jun 27.
Clinical trials, especially the randomized clinical trial, have been and will remain the gold standard for the evaluation of new interventions, including pharmaceuticals, biologics, medical devices, procedures, or behavioral modifications. Despite more than five decades of experience, there are still challenges in their design, conduct, monitoring, and analyses. Some of these challenges remain and some are emerging, in part due to the progress in genomics and proteomics. These issues may be statistical, logistical, or a combination. Included are follow-up of subjects who withdraw from intervention, the proposed use of recent adaptive designs, implementing noninferiority designs, reliance on surrogate markers, and gene transfer studies. Comparative effectiveness studies are of increasing interest but present major design and analysis issues. Forces external to the trial are also becoming more common.
临床试验，尤其是随机临床试验，一直以来并且仍将是评估新干预措施（包括药物、生物制剂、医疗器械、程序或行为改变）的金标准。尽管已经有超过五十年的经验，但在设计、实施、监测和分析方面仍然存在挑战。其中一些挑战仍然存在，而另一些则是新出现的，部分原因是基因组学和蛋白质组学的进步。这些问题可能是统计学的、后勤的或两者的组合。其中包括对退出干预的受试者的随访、最近自适应设计的拟议使用、实施非劣效性设计、依赖替代标志物和基因转移研究。比较有效性研究越来越受到关注，但也提出了重大的设计和分析问题。试验以外的力量也变得越来越常见。
