INFORMATION BASED DESIGN AND INTERIM MONITORING SOFTWARE

在许多临床实验中，研究人员随着时间的推移收集数据，以检验一个特定的研究假设。出于伦理和实际原因，研究者可能希望在中期分析时间点监测数据，以评估是否有足够的证据来拒绝或接受研究假设，如果存在这样的证据，则尽早终止研究。这样的实验被称为组序研究。总体目标是基于效率和统计信息的概念，在一个统一的框架内，开发一个非常通用的Microsoft Windows软件包，用于设计和中期监测组序临床试验。该软件将：*执行从几乎任何类型的数学模型，包括整个家族的广义线性模型，生存模型，线性和非线性混合效应模型，纵向研究半参数模型的兴趣参数组序列推理；*接口与市售的设计包，使他们能够将固定样本的研究转化为群体顺序的；*允许在不改变研究的操作特征的情况下，在每个临时监测时间点对投入研究的物理资源进行适应性重新设计。不存在具有所有这些功能的其他设计或临时监控软件。建议的商业应用：现在普遍的做法是，为可能的早期停止而设计大规模临床试验，并由独立的数据监测委员会进行监测。这对于由国家癌症研究所赞助的校外试验是强制性的。越来越多的监督委员会正在成立，以监督个别制药公司进行的关键癌症临床试验。设计和监控这类研究的软件很少，而及早停止研究的经济和伦理利益是巨大的，无论是为了有效还是徒劳。

In many clinical experiments, investigators collect data over time in an effort to test a particular research hypothesis. For ethical as well as practical reasons, the investigators might wish to monitor the data at interim analysis time-points to assess whether there is enough evidence to either reject or accept the research hypothesis and to terminate the study early if such evidence exists. Such an experiment is referred to as a group- sequential study. The overall goal is to develop a very general Microsoft Windows package for the design and interim monitoring of group- sequential clinical trials, within a single unified framework, based on the concepts of efficiency and statistical information. The software will: * perform group-sequential inference for the parameter of interest from practically any type of mathematical model including the entire family of generalized linear models, survival models, linear and non-linear mixed effects models, and semi-parametric models for longitudinal studies; * interface with commercially available design packages enabling them to convert fixed-sample studies into group-sequential ones; * permit adaptive redesign of the physical resources committed to a study at each interim monitoring-time point without altering the operating characteristics of the study. There exists no other design or interim monitoring software with all these capabilities. PROPOSED COMMERCIAL APPLICATIONS: It is now common practice for large scale clinical trials to be designed for possible early stopping and be monitored by independent data monitoring boards. This is mandatory for extramural trials sponsored by the National Cancer Institute. Increasingly monitoring boards are being established for pivotal cancer clinical trials conducted by individual pharmaceutical companies as well. Software to design and monitor such studies is scarce, while the economic and ethical benefits of stopping early, either for efficacy or futility, are enormous.