Simplified search strategies were effective in identifying clinical trials of pharmaceuticals and physical modalities

Background and Objectives: Assess the efficacy of simplified search strategies and identify the best electronic bibliographic database for clinical trials in the field of musculoskeletal disorders and pain. Methods: Clinical trials within selected reviews from the Cochrane Back, Musculoskeletal, and PaPaS Review Groups were searched using MEDLINE, EMBASE, CINAHL, and CENTRAL to identify which database included the highest percentage of trials. Simplified search strategies for each review were devised and compared to the original, more complex strategy for sensitivity, specificity and precision. Results: Individually, MEDLINE, and EMBASE included 90 and 89% of the relevant studies respectively, and 94% when combined. CENTRAL contained 87% and CINAHL 31%. Generally, simplified search strategies (two to four lines) had higher specificity than the original strategies (approximately 27 lines). Sensitivity was also high, but varied according to intervention. Super simple search strategies (one to two lines) proved as sensitive, but were slightly less specific, depending on the intervention. Both simple and super simple search strategies were often more precise than the original. Conclusion: Simplified search strategies are an effective, efficient way to search for clinical trials. They work best when the intervention is a pharmaceutical or a well-defined physical treatment. Their sensitivity, however, is not adequate for conducting systematic reviews. (c) 2005 Elsevier Inc. All rights reserved.