Pillastrini Paolo, Vanti Carla, Curti Stefania, Mattioli Stefano, Ferrari Silvano, Violante Francesco Saverio, Guccione Andrew
Associate Professor, Department of Biomedical and Neuromotor Sciences, University of Bologna, Italy.
Adjunct Professor Manual Therapy, Department of Biomedical and Neuromotor Sciences, University of Bologna, Italy.
J Manipulative Physiol Ther. 2015 Feb;38(2):159-66. doi: 10.1016/j.jmpt.2014.11.005. Epub 2014 Dec 12.
The aim of this study was to construct PubMed search strings that could efficiently retrieve studies on manual therapy (MT), especially for time-constrained clinicians.
Our experts chose 11 Medical Subject Heading terms describing MT along with 84 additional potential terms. For each term that was able to retrieve more than 100 abstracts, we systematically extracted a sample of abstracts from which we estimated the proportion of studies potentially relevant to MT. We then constructed 2 search strings: 1 narrow (threshold of pertinent articles ≥40%) and 1 expanded (including all terms for which a proportion had been calculated). We tested these search strings against articles on 2 conditions relevant to MT (thoracic and temporomandibular pain). We calculated the number of abstracts needed to read (NNR) to identify 1 potentially pertinent article in the context of these conditions. Finally, we evaluated the efficiency of the proposed PubMed search strings to identify relevant articles included in a systematic review on spinal manipulative therapy for chronic low back pain.
Fifty-five search terms were able to extract more than 100 citations. The NNR to find 1 potentially pertinent article using the narrow string was 1.2 for thoracic pain and 1.3 for temporomandibular pain, and the NNR for the expanded string was 1.9 and 1.6, respectively. The narrow search strategy retrieved all the randomized controlled trials included in the systematic review selected for comparison.
The proposed PubMed search strings may help health care professionals locate potentially pertinent articles and review a large number of MT studies efficiently to better implement evidence-based practice.
本研究旨在构建能够有效检索手法治疗（MT）相关研究的PubMed搜索字符串，特别是为时间紧迫的临床医生提供便利。
我们的专家选择了11个描述MT的医学主题词以及另外84个潜在主题词。对于每个能检索到100多篇摘要的主题词，我们系统地抽取了一部分摘要样本，据此估算与MT潜在相关的研究比例。然后我们构建了2个搜索字符串：1个狭义的（相关文章阈值≥40%）和1个广义的（包括所有已计算比例的主题词）。我们针对与MT相关的2种情况（胸痛和颞下颌关节疼痛）的文章测试了这些搜索字符串。我们计算了在这些情况下识别1篇潜在相关文章所需阅读的摘要数量（NNR）。最后，我们评估了所提出的PubMed搜索字符串识别纳入慢性下腰痛脊柱推拿治疗系统评价中的相关文章的效率。
55个搜索词能够提取100多条引用。使用狭义字符串找到1篇潜在相关文章的NNR，胸痛为1.2，颞下颌关节疼痛为1.3，广义字符串的NNR分别为1.9和1.6。狭义搜索策略检索到了所选用于比较的系统评价中包含的所有随机对照试验。
所提出的PubMed搜索字符串可能有助于医疗保健专业人员找到潜在相关文章，并高效地回顾大量MT研究，以更好地实施循证实践。
