Today I attended a fascinating lecture at Warwick Medical School by Andrew Booth of ScHARR on the subject of literature search futility. It is generally thought that in systematic reviews, the more databases that are searched the better. The result is that a lot of additional work as references are waded through. Booth analysed a broad selection of systematic reviews (n=50) and found that 85% of the papers they used were in Medline, and 38% of the reviews ony used papers that were in Medline. This raises questions of whether it is worth searching additional databases: is this bibliographic futility? Booth recommends the following as a protocol:
- Three core databases
- Follow up references
- Specialist databases
- Supplementary searches (hand searching, Google Scholar etc.)
I have only worked on one systematic review and we probably searched more databases than we needed to. Interestingly Booth recommends searching as more of an iterative approach (trying different terms) than as a single event. This is often how my own searches for library users turn out, especially if it is a tricky one.
A provocative and stimulating lecture!