Wilczynski Nancy L, Haynes R Brian
Department of Clinical Epidemiology and Biostatistics, McMaster University, Health Sciences Centre, 1200 Main Street West, Hamilton, Ontario, L8N 3Z5, Canada.
Ann Gen Psychiatry. 2006 Mar 23;5:4. doi: 10.1186/1744-859X-5-4.
General practitioners, mental health practitioners, and researchers wishing to retrieve the best current research evidence in the content area of mental health may have a difficult time when searching large electronic databases such as MEDLINE. When MEDLINE is searched unaided, key articles are often missed while retrieving many articles that are irrelevant to the search. The objectives of this study were to develop optimal search strategies to detect articles with mental health content and to determine the effect of combining mental health content search strategies with methodologic search strategies calibrated to detect the best studies of treatment.
An analytic survey was conducted, comparing hand searches of 29 journals with retrievals from MEDLINE for 3,395 candidate search terms and 11,317 combinations. The sensitivity, specificity, precision, and accuracy of the search strategies were calculated.
3,277 (26.8%) of the 12,233 articles classified in the 29 journals were considered to be of interest to the discipline area of mental health. Search term combinations reached peak sensitivities of 98.4% with specificity at 50.0%, whereas combinations of search terms to optimize specificity reached peak specificities of 97.1% with sensitivity at 51.7%. Combining content search strategies with methodologic search strategies for treatment led to improved precision: substantive decreases in the number of articles that needed to be sorted through in order to find target articles.
Empirically derived search strategies can achieve high sensitivity and specificity for retrieving mental health content from MEDLINE. Combining content search strategies with methodologic search strategies led to more precise searches.
希望检索心理健康领域当前最佳研究证据的全科医生、心理健康从业者和研究人员在搜索诸如MEDLINE等大型电子数据库时可能会遇到困难。在没有辅助的情况下搜索MEDLINE时，关键文章常常被遗漏，同时会检索到许多与搜索无关的文章。本研究的目的是制定最佳搜索策略以检测包含心理健康内容的文章，并确定将心理健康内容搜索策略与经校准以检测最佳治疗研究的方法学搜索策略相结合的效果。
进行了一项分析性调查，将29种期刊的手工检索结果与从MEDLINE中检索3395个候选搜索词和11317种组合的结果进行比较。计算了搜索策略的敏感性、特异性、精确性和准确性。
在29种期刊分类的12233篇文章中，有3277篇（26.8%）被认为与心理健康学科领域相关。搜索词组合的敏感性峰值达到98.4%，特异性为50.0%，而优化特异性的搜索词组合的特异性峰值达到97.1%，敏感性为51.7%。将内容搜索策略与治疗方法学搜索策略相结合可提高精确性：为找到目标文章而需要筛选的文章数量大幅减少。
通过实证得出的搜索策略在从MEDLINE中检索心理健康内容时可实现高敏感性和特异性。将内容搜索策略与方法学搜索策略相结合可实现更精确的搜索。
