Hennig Stefanie, Nyberg Joakim, Hooker Andrew C, Karlsson Mats O
Division of Pharmacokinetics and Drug Therapy, Department of Pharmaceutical Biosciences, Uppsala University, Box 591, SE 751 24 Uppsala, Sweden.
J Clin Pharmacol. 2009 Mar;49(3):323-35. doi: 10.1177/0091270008329560.
Optimal design has been used in the past mainly to optimize sampling schedules for clinical trials. Optimization on design variables other than sampling times has been published in the literature only once before. This study shows, as an example, optimization on the length of treatment periods to obtain reliable estimates of drug effects on longterm disease progression studies. Disease progression studies are high in cost, effort, and time; therefore, optimization of treatment length is highly recommended to avoid failure or loss of information. Results are provided for different drug effects (eg, protective and symptomatic) and for different lengths of studies and sampling schedules. The merits of extending the total study length versus inclusion of more samples per participants are investigated. The authors demonstrate that if no observations are taken during the washout period, a trial can lose up to 40% of its efficiency. Furthermore, when optimization of treatment length is performed using multiple possible drug effect models simultaneously, these studies show high power in discriminating between different drug effect models.
过去，最优设计主要用于优化临床试验的抽样计划。除抽样时间外，针对设计变量的优化此前仅在文献中出现过一次。本研究举例说明了对治疗期长度进行优化，以在长期疾病进展研究中获得药物效果的可靠估计。疾病进展研究成本高、工作量大且耗时久；因此，强烈建议优化治疗时长，以避免失败或信息丢失。针对不同的药物效果（如保护性和症状性）以及不同的研究时长和抽样计划给出了结果。研究了延长总研究时长与增加每位参与者样本量的优缺点。作者表明，如果在洗脱期不进行观察，试验效率可能会损失高达40%。此外，当同时使用多种可能的药物效果模型对治疗时长进行优化时，这些研究在区分不同药物效果模型方面显示出强大的功效。
