PubMed had a higher sensitivity than Ovid-MEDLINE in the search for systematic reviews

Objective: To compare the performance of Ovid-MEDLINE vs. PubMed for identifying randomized controlled trials of methotrexate (MTX) in patients with rheumatoid arthritis (RA). Study Design and Setting: We created search strategies for Ovid-MEDLINE and PubMed for a systematic review of MTX in RA. Their performance was evaluated using sensitivity, precision, and number needed to read (NNR). Results: Comparing searches in Ovid-MEDLINE vs. PubMed, PubMed retrieved more citations overall than Ovid-MEDLINE; however, of the 20 citations that met eligibility criteria for the review, Ovid-MEDLINE retrieved 17 and PubMed 18. The sensitivity was 85% for Ovid-MEDLINE vs. 90% for PubMed, whereas the precision and NNR were comparable (precision: 0.881% for Ovid-MEDLINE vs. 0.884% for PubMed and NNR: 114 for Ovid-MEDLINE vs. 113 for PubMed). Conclusion: In systematic reviews of RA, PubMed has higher sensitivity than Ovid-MEDLINE with comparable precision and NNR. This study highlights the importance of well-designed database-specific search strategies. (C) 2011 Elsevier Inc. All rights reserved.