OF RESEARCH PLAN: State the application's long-term objectives and specific aims, making reference to the health-relatedness of the project, describe concisely the methodology for achieving these goals, and discuss the potential of the research for technological innovation and commercial application. Avoid summaries of past accomplishments and the use of the first person. The abstract is meant to serve as a succinct and accurate description of the proposed work when separated from the application. Since abstracts of funded applications may be published by the Federal Government, do not include proprietary information. DO NOT EXCEED 200 WORDS. Large multi-institutional two-arm randomized clinical trials of time of failure (such as death from Cancer, or AIDS), are generally conducted as fixed sample studies. There are, however, ethical and economic pressures within the National Institutes of Health to monitor the accumulating data and to terminate such studies earlier than was originally planned, when strong evidence emerges that one treatment is clearly superior to the other, or that both treatments are equally ineffective. Recently developed group sequential statistical methods provide the flexibility to perform interim analyses on accumulating failure-time data, and stop early if warranted, while preserving the power and statistical significance of the corresponding fixed sample study. In Phase 1, Cytel took advantage of these new methodological advances and developed EAST, a prototype of a software package for planning and analysing group sequential clinical trials when the response is delayed and the data are monitored in groups. EAST was reviewed very favorably by a large number of prominant clinical trials statisticians. Phase 2 will build on the Phase 1 effort. East will be rewritten as a commercial statistical package. Its design module will provide one and two sided tests with lower and upper stopping boundaries in the power family, nad complete flexibility in the choice of the number of interim looks, the statistical power and the significance level. Its analysis module will exploit use function methodology for flexibility in monitoring accumulating data at previously unspecified time points, and will also provide stochastic curtailment capabilities for fixed sample studies, inference capabilities following early stopping, and exact permutational stopping boundaries for small sample problems. There will be powerful spreadsheet based user interface, with graphics support, similar to what was developed for the prototype. The package will be written in C, and will run on MS DOS machines, Unix work Stations, and Unix multi-user platforms.
