Lehmann H P, Nguyen B
Department of Medical Informatics, Johns Hopkins University School of Medicine, Baltimore, MD 21287-4461, USA.
MD Comput. 1997 Sep-Oct;14(5):353-9.
The World Wide Web provides a unique opportunity to reconsider how the results of scientific studies can best be presented to clinicians. For decades statisticians, philosophers, medical investigators, and others interested in data analysis have assumed that the Bayesian paradigm is the proper approach for reporting the findings of scientific analyses for use by client computers and readers. At the heart of that approach is the inclusion of the reader's preexisting knowledge and beliefs. Yet, to date, the methods for inclusion have been too complicated for non-statisticians to use. We believe that the World Wide Web provides an ideal environment for putting the Bayesian paradigm into practice: the author publishes the data from the server side, the reader uses the client to represent her or his prior belief, a downloaded program (a Java applet) combines the two. This article describes a prototype implementation for two-arm clinical trials with normally distributed outcomes.
万维网提供了一个独特的契机，让我们重新思考如何能将科学研究成果以最佳方式呈现给临床医生。几十年来，统计学家、哲学家、医学研究人员以及其他对数据分析感兴趣的人一直认为，贝叶斯范式是报告科学分析结果以供客户端计算机和读者使用的恰当方法。该方法的核心在于纳入读者先前已有的知识和信念。然而，迄今为止，纳入这些知识和信念的方法对于非统计学家来说过于复杂，难以使用。我们认为万维网为将贝叶斯范式付诸实践提供了理想的环境：作者从服务器端发布数据，读者使用客户端来表示其先验信念，一个下载的程序（Java小程序）将两者结合起来。本文描述了针对具有正态分布结果的双臂临床试验的原型实现。
