Papadopoulos M A
Department of Orthodontics, School of Dentistry, Aristotle University of Thessaloniki, Greece.
Orthod Craniofac Res. 2003 May;6(2):112-26.
In the last 10 years, health sciences literature has seen a significant increase in papers that report findings of meta-analyses. Meta-analysis overcomes the weakness of conventional narrative literature review. Moreover, it has established itself as a significant tool in clinical research and evidence-based medicine. But despite its station in the field, it is not without controversy. The aim of this paper is to present the main controversies that surround meta-analyses and to discuss the most important issues of conduct and reporting of results derived from these studies.
Controversial issues include the ability of the investigators to combine studies that differ in important aspects; such as study populations, experimental designs and quality controls, and the like. It has been argued that selection bias could exist. Thus, it is common for the results of metaanalyses to be contradictory. As any other statistical procedure or analytical approach, meta-analyses too can be misused or abused. Misleading results could be avoided if certain basic principles are followed in the conduct of a meta-analysis. In particular, attention should be given to: (a) the methods of literature search for the studies to be included in the analysis, (b) the measures taken to reduce or eliminate bias, (c) the inclusion and exclusion criteria of the individual studies from the analysis, (d) how the data will be extracted from the literature, and (e) how the data will be analyzed statistically.
Reports of meta-analytical studies carried out with the above parameters in mind yield a more objective appraisal of the evidence, and a more precise estimate of the treatment effect.
在过去10年中，健康科学文献中报告荟萃分析结果的论文数量显著增加。荟萃分析克服了传统叙述性文献综述的弱点。此外，它已成为临床研究和循证医学中的一项重要工具。但尽管它在该领域占据一席之地，却并非毫无争议。本文旨在介绍围绕荟萃分析的主要争议，并讨论这些研究在实施和结果报告方面的最重要问题。
有争议的问题包括研究人员合并在重要方面存在差异的研究的能力，例如研究人群、实验设计和质量控制等。有人认为可能存在选择偏倚。因此，荟萃分析的结果相互矛盾是很常见的。与任何其他统计程序或分析方法一样，荟萃分析也可能被滥用或误用。如果在进行荟萃分析时遵循某些基本原则，就可以避免产生误导性结果。特别是，应注意：(a) 分析中纳入研究的文献检索方法；(b) 为减少或消除偏倚采取的措施；(c) 分析中各个研究的纳入和排除标准；(d) 如何从文献中提取数据；以及 (e) 如何对数据进行统计分析。
在考虑上述参数的情况下进行的荟萃分析研究报告，能对证据进行更客观的评估，并对治疗效果进行更精确的估计。
