DESIGN POWER SOFTWARE FOR CLINICAL STUDIES

此第一阶段SBIR建议涉及用于计算样本量和 生物医学研究设计的统计力量。对于第一阶段，我们的 计划将实现两个新的模拟设计的时间到 事件研究：Halpern和Brown[1987]的一种方法 生存数据，可能取自已发表的文章，以估计样本 一项新研究的规模和力量，以及我们为 这些累积的生存数据在每次更新时都会重新分析。第一阶段 还将包括重复测量的功率计算(纵向) 具有持续结果的研究。所有这些在当前版本中都不可用 设计软件。我们的第二阶段产品包将包括简单、图形化 探索在设计特征、计算和计算之间的权衡 一系列临床和流行病学设计以及对背景敏感的建议 给预期用户(例如，涉及研究设计或 统计学家)。最近对这一产品类别的评论表明 对全面、灵活的医疗设计软件的大量需求， 环境和工业研究，但目前的状况是 不能令人满意。我们已经研究了这些评论和软件，并相信 我们建议的产品更丰富、更全面、更全面 很有吸引力。

This Phase I SBIR proposal concerns PC software to compute sample size and statistical power for the design of biomedical studies. For Phase I, our program will implement two novel simulations for the design of time-to- event studies: a method of Halpern and Brown [1987] that uses existing survival data, perhaps taken from published articles, to estimate sample size and power for a new study, and a method we developed for studies in which accumulating survival data are re-analyzed at each update. Phase I will also include a power calculation for repeated measures (longitudinal) studies with continuous outcomes. None of these is available in current design software. Our Phase II package will include easy, graphical exploration of tradeoffs between design features, calculations for a wide range of clinical and epidemiological designs and context-sensitive advice to the intended user (e.g., physician concerned with study design or statistician). Recent reviews of this product category suggest a substantial demand for comprehensive, flexible design software for medical, environmental and industrial studies but that the current state is unsatisfactory. We have studied these reviews and the software and believe that our proposed product is more informative, more comprehensive, and more attractive.