Tutorial: survival analysis--a statistic for clinical, efficacy, and theoretical applications.
Tutorial: survival analysis--a statistic for clinical, efficacy, and theoretical applications.
复制标题
教程：生存分析——临床、疗效和理论应用的统计。
DOI：
10.1044/jslhr.4202.432
复制
发表时间：
1999
期刊：
影响因子：
--
通讯作者：
Gruber,FA
中科院分区：
文献类型：
--
作者：
Gruber,FA
当前要求增加对治疗效果和效率的研究关注，以及对改进的发展模型的要求，要求对纵向结果数据进行分析。纵向语音和语言数据的统计处理是困难的，但在医学、精算、制造和社会学中有一系列常用的统计技术尚未在语音或语言研究中使用。生存分析作为一种方法被引入，它避免了许多其他技术的统计问题，因为它将时间视为结果。在生存分析中，不仅计算群体的概率，也计算群体中个人的概率。这是临床工作的一大优势。本文以语音输出为例，对非参数和半参数生存分析进行了基本的介绍。还简要讨论了精算分析和临床直觉之间的潜在冲突。
Current demands for increased research attention to therapeutic efficacy, efficiency, and also for improved developmental models call for analysis of longitudinal outcome data. Statistical treatment of longitudinal speech and language data is difficult, but there is a family of statistical techniques in common use in medicine, actuarial science, manufacturing, and sociology that has not been used in speech or language research. Survival analysis is introduced as a method that avoids many of the statistical problems of other techniques because it treats time as the outcome. In survival analysis, probabilities are calculated not just for groups but also for individuals in a group. This is a major advantage for clinical work. This paper provides a basic introduction to nonparametric and semiparametric survival analysis using speech outcomes as examples. A brief discussion of potential conflicts between actuarial analysis and clinical intuition is also provided.