Can we prioritise which databases to search? A case study using a systematic review of frozen shoulder management

Background Systematic reviews risk producing biased conclusions if a comprehensive search to identify eligible studies is not undertaken, but little evidence exists to guide prioritisation of databases to search when resources are limited. Objectives A systematic review examining interventions for managing frozen shoulder (adhesive capsulitis) was used to investigate the performance of bibliographic databases in identifying the included studies, the smallest combination of databases required to retrieve all included studies, and the performance of the searches themselves. Methods We calculated the yield of included studies from each of 15 databases, and the recall and precision of each search strategy. We investigated differences between the presence of a record in a database and its retrieval. Results Thirty of 31 studies were present in at least one database. Yields of individual databases ranged from 0% to 90% (median 23%). Two combinations of databases identified all 30 studies: Cochrane Central Register of Controlled Trials (CENTRAL) and Science Citation Index (SCI); or CENTRAL, MEDLINE and PreMEDLINE. Conclusions In a systematic review of a range of interventions used to manage frozen shoulder, at least two databases and reference checking were required to retrieve all included studies, but searching for future reviews should not be restricted.