Article type
Year
Abstract
Introduction: The musculoskeletal group in the Cochrane Collaboration is carrying out a series of systematic reviews of the randomised controlled trials (RCTs) of interventions to prevent osteoporosis and osteoporotic fractures. The clinical epidemiology units at the Universities of Ottawa (CEUO) and Newcastle (CCEBN) are collaborating in a comparative analysis of the results of different search strategies.
Methods: Both groups have performed searches on interventions involving estrogen treatment, calcium supplementation, and calcitonin treatment. CCEBN searched the CD ROM Silver Platter Medline Database (1970-1993), online Medline (BRS; 1989-1993) and Excerpta Medica (1974-1993). Both MeSH headings and free text searching were used to identify articles involving osteoporosis, relevant treatment terms, and selected bone terms (including bone mass, bone mineral content, bone mineral density). Bibliographies of all articles were searched manually. The CEUO used a modified version of the Cochrane search strategy on Medline. Bibliographies of articles are currently being searched.
Results: CCEBN's broad search strategy resulted in the identification of several thousand titles that were reviewed, of which only a very small proportion met the criteria for RCTs. The more focussed strategy of CEUO identified several hundred articles. It appears that the CEUO search identified around 30% of the trials retrieved by CCEBN but this may increase after bibliography searching.
Discussion: The broad and labour intensive search is likely to have identified more of the relevant RCTs in this literature. A comparison of the yield rates and examination of the characteristics of trials, which were missed by each of the strategies, will lead to recommendations for a more efficient and effective search strategy.
Methods: Both groups have performed searches on interventions involving estrogen treatment, calcium supplementation, and calcitonin treatment. CCEBN searched the CD ROM Silver Platter Medline Database (1970-1993), online Medline (BRS; 1989-1993) and Excerpta Medica (1974-1993). Both MeSH headings and free text searching were used to identify articles involving osteoporosis, relevant treatment terms, and selected bone terms (including bone mass, bone mineral content, bone mineral density). Bibliographies of all articles were searched manually. The CEUO used a modified version of the Cochrane search strategy on Medline. Bibliographies of articles are currently being searched.
Results: CCEBN's broad search strategy resulted in the identification of several thousand titles that were reviewed, of which only a very small proportion met the criteria for RCTs. The more focussed strategy of CEUO identified several hundred articles. It appears that the CEUO search identified around 30% of the trials retrieved by CCEBN but this may increase after bibliography searching.
Discussion: The broad and labour intensive search is likely to have identified more of the relevant RCTs in this literature. A comparison of the yield rates and examination of the characteristics of trials, which were missed by each of the strategies, will lead to recommendations for a more efficient and effective search strategy.