New Challenges Facing Retrieval of Papers Published in China on Systematic Evaluation and Meta-Analysis of Intervention
葛龙,安妮,曾巧铃,马继春,毛婧,石新彤,徐俊峰,田金徽
DOI: https://doi.org/10.3969/j.issn.1671-3982.2013.05.001
2013-01-01
Abstract:Objective To analyze the papers published in China on evaluation and meta-analysis of tests and their screening methods.Methods Papers on evaluation and meta-analysis of tests published in China before 2011 and covered in CBM were retrieved using the retrieval words including diagnosis, specificity,sensitivity,systematic evaluation,meta-analysis,and systematic reviews.Their retrieved databases and retrieval strategies were analyzed using the SPSS 17.0.Results Of the 235 papers on evaluation and meta-analysis of tests that were included in this study,176(74.9%) retrieved databases both in Chinese and English,29(12.3%) retrieved databases in Chinese,28(11.9%) retrieved databases in English,139(67.8%) retrieved 2 or more databases in Chinese,151(74.0%) retrieved 2 or more databases in English.The most commonly retrieved databases in Chinese were CNKI,CBM,VIP,and Wanfang,the most frequently retrieved databases in English were PubMed,EMBASE,Cochrane Library,and OVID,whereas 47.2%,88.1%,46.4%,10.2%,and 8.9% of the papers on evaluation and meta-analysis of tests reported manual retrieval,retrieval start and end time,reference retrieval,contact with experts and authors,and search engine,respectively.Only 44.3% of the papers reported that target disease,tests to be evaluated and diagnostic accuracy were used as retrieval words in literature search and 4.4% did not report that they used any of the 3 retrieval words,48.9% reported their steps of screening papers,and 12.3% only provided a flow screening chart.Conclusion The databases have not covered all papers on evaluation and meta-analysis of tests and only a small number of papers on evaluation and meta-analysis of tests searched tests databases.Suggestions are thus put forward for those who are engaged in research of evaluation and meta-analysis of tests should retrieve at least 2 commonly used databases,perfect search engines,strengthen reference retrieval,and report their retrieval strategies and paper-screening methods.