As systematic reviews aim to be comprehensive in the evidence they review, multiple databases must be searched. For example the Cochrane Library recommends searching:
You may also want to include:
Note. Trial registries should be searched for unpublished trials to reduce influencing results by publication bias.
Searches should be saved for each database so that search strategies can be reproduced in the final report. Records should also be kept of when key searches were carried out.
See our Guide to literature searching for advice on searching.
In order to be as comprehensive as possible it is recommended that searches are as broad as possible with inclusion or exclusion decided at the screening stage. Tightly targeted searches risk failing to identify relevant papers.
Meta-evidence blog - Searching Grey Literature