摘要 : Evidence-based medicine has recently received a large amount of attention in medical research. To help clinical practices use evidence-based medicine, it should be easy to find the best current evidence that is relevant to the cli... 展开 Evidence-based medicine has recently received a large amount of attention in medical research. To help clinical practices use evidence-based medicine, it should be easy to find the best current evidence that is relevant to the clinical question and has high methodological quality. However, searching for relevant articles and appraising their validity is demanding work for most clinicians. We hypothesize that, through an effective design that addresses the two major aspects - relevance and quality - together with a ranking algorithm, search engines can automatically retrieve articles that are relevant to clinical questions and are based on valid evidence. The contribution of this study has two parts. First, we approach this problem by combining methodologies. After designing a suitable document query-relevance score and methodological quality score, we combined them using various fusion methods. The result was a twofold increase in the mean average precision. Second, for correct evaluation, we built a test collection using a preexisting reliable database, the Cochrane Reviews, which allowed robust and comprehensive evaluation. 收起