动物实验系统评价/Meta分析检索策略报告情况调查

The Search Strategy of Systematic Review/Meta-analysis of Animal Research: A Survey

目的对国内外动物实验系统评价/Meta分析(SR/MA)检索策略的全面性及报告现状进行调查分析,为今后动物实验SR/MA检索策略的制定和报告提供参考依据。方法计算机检索Ovid-MEDLINE、Ovid-EMbase、Ovid-BIOSIS Previews、CBM、CNKI、VIP和WanFang Data数据库,检索时间均从建库至2015年1月。纳入国内外医学相关的动物实验SR/MA,不限动物种属。采用自制的资料提取表提取纳入研究的基本特征、检索数据库、其他检索来源、检索策略等相关信息,采用百分比和频数进行统计描述。结果最终纳入181篇动物实验SR/MA。不足30%的SR/MA检索了3个以上的数据库,65.7%报告进行了补充检索。86.2%(156/181)的SR/SMA研究报告了检索词,而报告完整检索策略的仅有12个;33个SR/MA的检索词包含了具体的动物名称,27个SR/SMA使用了"animal/experimental model/"一词,20个使用了"limit to animal"的限定功能。71.3%的SR/MA报告了具体的检索时间范围,44.2%报告说明了是否对纳入研究进行语种限制,其中的23.8%限制语种,且多限定为英文。结论目前,国内外的动物实验SR/MA在数据库、检索词以及检索策略的选择、报道上仍存在一些问题,因此我们建议:1在数据库的选择上,需要根据研究领域的不同尽可能选择足够具有代表性的多个数据库,外文数据库建议至少检索MEDLINE/PubMed、EMbase、BIOSIS Preview,中文数据库建议至少检索CBM、CNKI和WanFang Data;2根据研究对象将具体动物种属名称和(或)"animal(动物)"作为检索词进行检索,并根据数据库的特点使用"limit to animal"功能;3动物实验SR/MA的检索策略报告应该包括检索来源、检索时间范围、语种限制、研究类型限制、检索词以及完整的检索策略。此外,需要报告检索人员的资质;4有关期刊需要将"完整地报告检索策略"引入稿约,以提高SR/SMA的透明度和可信度。

Objective To provide a reference for developing search strategy of systematic reviews/ meta-analyses（SRs/MAs） of animal researches（ARs） in future, we investigated and analyzed the search strategy of SRs/MAs of ARs at home and abroad. Method Ovid-MEDLINE, Ovid-EMbase, Ovid-BIOSIS previews, CBM, CNKI, VIP and Wan Fang Data were searched from inception to January 2015, to collect SRs/MAs of ARs that related to medicine. No limitation in species. Two reviewers independently screened literature, extracted basic characteristics, databases and other sources searched, characteristics of reported search strategy of included studies. Result A total of 181 SRs/MAs of ARs were finally included. Less than 30% SRs/MAs searched more than three databases, 65.7% reported supplementary retrieval. 86.2%（156/181） SRs/MAs reported search terms, but only 12 reported search strategy; the search terms of 33 studies included specific name of animal, 27 studies used ＂animal/experimental model/＂, 20 studies used ＂limit to animal＂. 71.3% SRs/MAs reported specific time limitation of searching, 44.2% reported whether limited language, 23.8% of them limited language, and more of these limited to English. Conclusion At present, there are still some problems in SRs/MAs of ARs at home and abroad when choice database, search terms and search strategy, so we advise that： 1） It’s necessary to choice typical databases as many as possible according to search field; 2） Using specific animal’s name and（or） ＂animal＂ as one of search terms, and using ＂limit to animal＂ according to characteristic of different databases; 3） The reporting of search strategy of SRs/MAs of ARs should include search sources, time limitation, language limitation, limitation of inclusion type, search terms and complete search strategy, besides, reporting knowledge of reviewers is also necessary; 4） To improve transparency and clarity of SRs/MAs of ARs, some related journals should introduce ＂reporting complete search strategy＂ in their instruction for authors.