Callcut Rachael A, Branson Richard D
Department of Surgery, University of Cincinnati, 231 Albert Sabin Way, Cincinnati, OH 45267-0558, USA.
Respir Care. 2009 Oct;54(10):1379-85.
Review papers commonly summarize the current knowledge on a selected topic. These types of papers are considered narrative reviews. Narrative reviews rarely detail the methods used to select the literature included, nor do the authors typically report the purpose of the review. Narrative reviews may be biased due to inadequate literature reviews or individual beliefs. A systematic review limits bias by disclosing the purpose of the paper, the assembly of the literature, and the appraisal of study quality. A meta-analysis, a specific style of systematic review, quantitatively pools data from individual studies for re-analysis. Pooling data increases the sample size and improves statistical power. The common representation of a meta-analysis is the forest plot. The forest plot demonstrates the odds ratio of individual studies, the weight each trial contributes to the analysis, and the 95% confidence intervals. Systematic reviews and meta-analyses are not without shortcomings, including issues related to study heterogeneity. Because of their transparency, systematic evaluations of the literature are superior to narrative reviews.
综述论文通常会总结关于某一选定主题的当前知识。这类论文被视为叙述性综述。叙述性综述很少详细说明用于选择所纳入文献的方法，作者通常也不会报告综述的目的。由于文献综述不充分或个人信念，叙述性综述可能存在偏差。系统综述通过披露论文的目的、文献的收集以及研究质量的评估来限制偏差。荟萃分析是系统综述的一种特定形式，它对来自各个研究的数据进行定量汇总以便重新分析。汇总数据会增加样本量并提高统计效力。荟萃分析的常见呈现形式是森林图。森林图展示了各个研究的比值比、每个试验对分析的贡献权重以及95%置信区间。系统综述和荟萃分析并非没有缺点，包括与研究异质性相关的问题。由于其透明度，对文献的系统评价优于叙述性综述。
