Article type
Year
Abstract
Introduction: Identification of relevant trials is a major undertaking when conducting a systematic review of an intervention. Electronic searches of medical databases such as MEDLINE or EMBASE can be sensitive, depending on the terms used but have very low specificity, resulting in a time-consuming low efficiency process. Ideally, searching a specialised register of controlled clinical trials (CCTs) should increase the specificity of the search with obvious advantages in relation to human and economic resources. The Cochrane Controlled Clinical Trials Register (CENTRAL) is such a register which includes CCTs of interventions in all medical fields.
Objectives: The objective of this study was to evaluate the performance of CENTRAL for the retrieval of controlled clinical trials (CCTs) in selected musculoskeletal disorders
Methods: We evaluated the performance of CENTRAL to identify CCTs relevant to the fields of low back pain, rheumatoid arthritis and osteoporosis. MeSH terms and textwords were used to retrieve relevant references We used the first issue of 1999. Two other databases were used as benchmarks: a) a database of CCTs from a previous study which compared the performance of MEDLINE and EMBASE. This database includes publications for 1988 and 1994 for the three musculoskeletal disorders using comprehensive electronic searches and hand searching of selected journals. Only a small proportion of the retrieved publications had been found to be CCTs; b) the second database includes CCTs of selected interventions for the treatment of rheumatoid arthritis. This database was compiled through comprehensive electronic searches and hand searching of the bibliographies of all the CCTs electronically retrieved.
Results: The databases of CCTs used as benchmark included most of the CCTs in CENTRAL. Yet, several CCTs in the databases were not in CENTRAL. We found several instances where publications in CENTRAL were not CCTs; nevertheless, as expected, the specificity for CCTs was very high, when compared to MEDLINE or EMBASE searches.
Discussion: Our findings suggest that as it currently stands, CENTRAL is very specific but not as sensitive for the retrieval of CCTs in selected medical fields. Use of other searching methods such as additional electronic searches or selected hand searches may still be required.
Objectives: The objective of this study was to evaluate the performance of CENTRAL for the retrieval of controlled clinical trials (CCTs) in selected musculoskeletal disorders
Methods: We evaluated the performance of CENTRAL to identify CCTs relevant to the fields of low back pain, rheumatoid arthritis and osteoporosis. MeSH terms and textwords were used to retrieve relevant references We used the first issue of 1999. Two other databases were used as benchmarks: a) a database of CCTs from a previous study which compared the performance of MEDLINE and EMBASE. This database includes publications for 1988 and 1994 for the three musculoskeletal disorders using comprehensive electronic searches and hand searching of selected journals. Only a small proportion of the retrieved publications had been found to be CCTs; b) the second database includes CCTs of selected interventions for the treatment of rheumatoid arthritis. This database was compiled through comprehensive electronic searches and hand searching of the bibliographies of all the CCTs electronically retrieved.
Results: The databases of CCTs used as benchmark included most of the CCTs in CENTRAL. Yet, several CCTs in the databases were not in CENTRAL. We found several instances where publications in CENTRAL were not CCTs; nevertheless, as expected, the specificity for CCTs was very high, when compared to MEDLINE or EMBASE searches.
Discussion: Our findings suggest that as it currently stands, CENTRAL is very specific but not as sensitive for the retrieval of CCTs in selected medical fields. Use of other searching methods such as additional electronic searches or selected hand searches may still be required.