David M. Crowther, UT MD Anderson Cancer Center, 1515 Holcombe Blvd, Unit 0377, Houston, TX 77030, USA.
Nutr Clin Pract. 2013 Aug;28(4):459-62. doi: 10.1177/0884533613490742. Epub 2013 Jun 6.
The purpose of this article is to discuss systematic reviews, how they are performed, and their associated strengths and limitations. A systematic review is an assessment of evidence involving exact methods to systematically identify, select, and critically evaluate all available literature on a particular topic. Unlike most narrative reviews, systematic reviews have defined methods established a priori for searching, evaluating, extracting, synthesizing, and reporting available evidence. Key characteristics differentiating systematic reviews from most narrative reviews include: clearly stated objectives, pre-defined inclusion/exclusion criteria, an explicit reproducible methodology, systematic exhaustive searches to identify all sources of evidence, an assessment of the validity for each included study, and a systematic presentation of the study characteristics/results. Though there are significant advantages to systematic reviews, there are also clear limitations such as: the quality of included evidence; heterogeneity and homogeneity of included studies; and publication bias. Even with these limitations, systematic reviews are beneficial to front line clinicians when the quantity of evidence is so substantial that reviewing and synthesizing it is not feasible, available evidence is conflicting, or when the robustness of available evidence is unknown.
本文旨在讨论系统评价，包括其实施方法以及相关的优势和局限性。系统评价是对证据的评估，涉及到确切的方法，以系统地识别、选择和批判性评估特定主题的所有现有文献。与大多数叙述性综述不同，系统评价具有预先确定的用于搜索、评估、提取、综合和报告现有证据的方法。将系统评价与大多数叙述性综述区分开来的关键特征包括：明确的目标、预先确定的纳入/排除标准、明确的可重复的方法、系统地进行全面搜索以确定所有证据来源、对每个纳入研究的有效性进行评估，以及系统地呈现研究特征/结果。尽管系统评价有显著的优势，但也有明显的局限性，如纳入证据的质量、纳入研究的异质性和同质性以及发表偏倚。即使存在这些局限性，当证据数量非常大以至于审查和综合不可行、现有证据存在冲突，或者现有证据的稳健性未知时，系统评价对一线临床医生仍然是有益的。
