ADAPTATION OF NEW STATISTICAL IDEAS FOR MEDICINE

新的医学统计理念的适应

• 批准号: 3203151
• 负责人: BRADLEY EFRON
• 金额: $ 17.06万
• 依托单位: STANFORD UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 1993
• 资助国家: 美国
• 起止时间: 1993-01-15 至 1995-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/3203151
• 关键词: automated medical record system; biohazard control; computer center; computer simulation; data collection methodology /evaluation; epidemiology; health science research; health surveys; human data; human therapy evaluation; imaging /visualization /scanning; information systems; mathematics; neoplasm /cancer therapy; statistics /biometry; vital statistics

The long-term goal of this research grant is to bring new developments in statistical theory and methodology to bear on the practical problems of biostatistics and medicine. The researchers pursue this goal from their joint positions in the Stanford medical school, where they work on specific biostatistical projects, and the development and review. Research on the grant involves the creation of new statistical methods, study of new and existing techniques through mathematical analysis and computer simulation, and trial applications of these procedures to ongoing research projects at the medical school. The following five areas will be the focus of the research: (1) bootstrap methods, (2) recursive partitioning and classification trees, (3) modern regression methods and their extensions, (4) image reconstruction methods, (5) hazard rates and censored data. Successful new statistical methodology finds application across a wide spectrum of biomedical applications. On the basis of past experience, the topics studied here will be particularly useful in cancer therapy clinical trials, cardiovascular research, and areas like gait analysis where diagnostic measures tend to have very complicated structure.

这项研究资助的长期目标是为以下领域带来新的发展： 统计理论和方法，以承担的实际问题， 生物统计学和医学。 研究人员追求这一目标， 在斯坦福大学医学院的联合职位，他们在那里工作的具体 生物统计项目，以及发展和审查。 研究 赠款涉及创造新的统计方法，研究新的和 现有技术通过数学分析和计算机模拟， 并将这些程序试用于正在进行的研究项目， 医学院 以下五个领域将是 研究：（1）引导方法，（2）递归划分和 分类树，（3）现代回归方法及其扩展， (4)图像重建方法;（5）风险率和删失数据。 成功的新统计方法适用于广泛的 一系列生物医学应用。 根据以往的经验， 本文研究的主题在癌症治疗临床中特别有用。 试验，心血管研究，以及步态分析等领域， 诊断措施往往具有非常复杂的结构。