統合分析－（I）基本統計方法

Statistical Methods in Meta-Analysis

葉明功(Ming-Kung Yeh) ； 白璐(Lu Pai)

《Journal of Medical Sciences》 18卷5期 (1998/04) Pp. 273-296

摘要

統合分析是以系統性量化，結合資料，以達到具體客觀的評述。首先藉由系統化的定義問題，設定納入或排除條件，依次萃取資料，而後經由統計方法結合相關資料，以得到具體結論。其中統計方法的運用在統合分析研究中，佔有重要地位，可用於整合不同研究中相關之資料，比較差異，及求出信賴區間。本章將詳述各種常用之統計方法，以達上述之目標。包括二項式數據（以相對危險度、優勢比、率差比較）、順序數據，及累計統合分析，並以實例說明並印證之。

關鍵字

統合分析 ； 統計

並列摘要

Meta-analysis is a quantitative approach for systematically combining the results of previous research in order to arrive at conclusions about the body of research. Studies of a topic are first systematically identified. Criteria for including and excluding studies are defined, and data from the eligible studies are abstracted. Last, the data are combined statistically, yielding a quantitative estimate of the size of the effect of treatment and a test of homogeneity in the estimate of effect size. One of the key point is the statistical methods in meta-analysis. The goals of the statistical analysis of data from several studies are to estimate a summary measure of effect size, the variance of the summary estimate of effect size, and a confidence interval. This chapter describes the statistical methods that address these three goals. Section 1 discusses binary data using the relative risk, odds ratio and rate difference. Section 2 discusses the ordinal data, also shows some examples. Sections 3 discusses the cumulative meta-analysis.

並列關鍵字

meta-analysis ； statistics