Walker-Dilks Cindy, Wilczynski Nancy L, Haynes R Brian
Health Information Research Unit, Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada.
Appl Nurs Res. 2008 May;21(2):98-103. doi: 10.1016/j.apnr.2006.04.004.
We developed search strategies for detecting sound articles on causation and prognosis in Cumulative Index to Nursing and Allied Health Literature (CINAHL) in the year 2000. An analytic survey was conducted, comparing hand searches of 75 journals with retrievals from CINAHL for 5,020 search terms and 11,784 combinations for causation and 9,946 combinations for prognosis. For detecting sound causation studies, a three-term strategy maximized sensitivity at 97.0% with a specificity of 52.3%. For detecting sound prognosis studies, a three-term strategy maximized sensitivity at 92.2% with a specificity of 50.0%. These search filters will enhance the searching efforts of clinicians and researchers.
2000年，我们制定了检索策略，以在《护理学与健康相关学科累积索引》（CINAHL）中查找关于因果关系和预后的可靠文章。我们进行了一项分析性调查，将对75种期刊的人工检索结果与CINAHL中5020个检索词、11784种因果关系组合以及9946种预后组合的检索结果进行比较。对于检测可靠的因果关系研究，一个三项策略的灵敏度最高可达97.0%，特异度为52.3%。对于检测可靠的预后研究，一个三项策略的灵敏度最高可达92.2%，特异度为50.0%。这些检索过滤器将提高临床医生和研究人员的检索效率。
