Observational studies provide important evidence for clinicians and policy makers to support their decision making. Systematic reviews of observational studies help answer research questions focusing on prognosis, cause, risk factors, and complications. However, when undertaking searches finding observational studies remains challenging because of the poor and inconsistent indexing in literature databases. Two main literature databases in healthcare are MEDLINE and Embase. To enable more effective and efficient retrieval of a particular type of study, literature search strategies can be complemented with methodological filters, that is, a collection of terms relating to a particular study design. The performance of a search strategy is normally evaluated by sensitivity (i.e., an ability to retrieve all relevant studies) and precision (i.e., an ability to retrieve relevant studies). In this Cochrane review, we aimed to collate evidence on performance of methodological filters aspart of search strategies for identifyingobservational studies inMEDLINE and Embase. We searched MEDLINE (1946-), Embase (1974-), CINAHL (1937-), the Cochrane Library (1992-), Google Scholar, and Open Grey in April 2018 for studies comparing sensitivity or precision of a search strategy containing a methodological filter to identify observational studies in MEDLINEandEmbaseagainst a reference standard.We included studies using the relative recall approach for the validation and assessment of the methodological filter performance. Relative recall approach compares the performance of a search strategy with and without a methodological filter.We also included studies that compared the performance of two or more different methodological filters. In addition, we scanned the reference lists of relevant articles for additional eligible studies. Two review authors independently screened articles, extracted relevant information, and assessed the quality of the search strategies using the InterTASC Information Specialists’ Sub-Group (ISSG) Search Filter Appraisal Checklist. We found two eligible studies reporting on 18 methodological filters, comprising six MEDLINE, six Embase, and six combined MEDLINE/Embase filters. The methodological filters evaluated in the included studies were developed based on the terms retrieved from the reference standard studies. The first study evaluated six filters for retrieval of observational studies of surgical interventions. These filters were a Precision Terms Filter (comprising terms with higher precision while maximum sensitivity was maintained) and a Specificity Terms Filter (comprising terms with higher specificity while maximum

中文翻译：

---

Cochrane评论：搜索策略，以识别MEDLINE和EMBASE中的观察性研究。

观察性研究为临床医生和政策制定者支持其决策提供了重要证据。观察性研究的系统评价有助于回答有关预后，病因，危险因素和并发症的研究问题。但是，由于文献数据库中的索引不足且不一致，因此进行搜索时发现观察性研究仍然具有挑战性。医疗保健领域的两个主要文献数据库是MEDLINE和Embase。为了能够更有效地检索特定类型的研究，可以使用方法过滤器（即与特定研究设计相关的术语集合）来补充文献检索策略。搜索策略的效果通常通过敏感度（即检索所有相关研究的能力）和精确度（即 检索相关研究的能力）。在此Cochrane综述中，我们旨在收集方法筛选器性能的证据，作为用于鉴定MEDLINE和Embase中观测研究的搜索策略的一部分。我们在2018年4月对MEDLINE（1946-），Embase（1974-），CINAHL（1937-），Cochrane图书馆（1992-），Google Scholar和Open Gray进行了搜索，以比较包含方法论的搜索策略的敏感性或准确性过滤器以识别MEDLINE和Embase中参考标准的观察性研究。我们纳入了使用相对召回方法进行过滤性方法学性能验证和评估的研究。相对召回方法比较了使用和不使用方法过滤器的搜索策略的效果。我们还进行了比较两个或多个不同方法过滤器性能的研究。此外，我们还扫描了相关文章的参考列表，以进行其他符合条件的研究。两位评论作者独立筛选了文章，提取了相关信息，并使用InterTASC信息专家小组（ISSG）搜索过滤器评估清单评估了搜索策略的质量。我们发现两项合格的研究报告了18种方法学过滤器，包括6种MEDLINE，6种Embase和6种组合MEDLINE / Embase过滤器。根据从参考标准研究中检索到的术语，开发了纳入研究中评估的方法学过滤器。第一项研究评估了六个过滤器，以检索外科手术干预的观察性研究。