If you are going to search multiple databases, you will need to convert or 'translate' your search to the next database. The more concise your search is, the easier it will be to convert it. Avoid overlap in index terms e.g. MeSH or Emtree.
For instance these MeSH terms:
"General Practice"[Mesh](~74.000 items) > "Family Practice"[Mesh](~65.000 items)
The broader term "General Practice" finds around 74k items and includes all of the items found using the more narrow term "Family Practice". No need to use both, for a broad search use "General Practice".
When searching the title and abstract: avoid overlap between broader and narrower terms as well. The single word term:
trials[Title/Abstract] will find all trials, including clinical trials[Title/Abstract].
The two word term clinical trials[Title/Abstract] will find all clinical trials, so both randomized clinical trials[Title/Abstract] as well as randomised clinical trials[Title/Abstract] will be found. To include both singular as well as the plural you can add an asterisk, or wildcard: clinical trial*[Title/Abstract].
Determine how broad your search should be and and choose an option. Evaluate your search well in the database where you start, before converting it to the next one, which avoids that you have to make changes in multiple searches.
It's recommended to use a combination of text terms and indexed key words, MeSH terms for PubMed. When using the 'Advanced Search' the default field is selected as 'All Fields'. When combining searches from the Search History that use MeSH or Title/Abstract, it's not needed to change the 'All Fields'. PubMed prioritises whatever is in [square brackets] over the default 'All Fields'. See below ↓
If you plan to search multiple databases, as is common for a systematic review, make sure that your search is well structured, and has no overlapping terms, before you translate it into another database. Translating to the next database is not something you want to do more than once.
Snowballing - Citation searching
Snowballing has two goals, first of all it can be a check to see whether or not your current search is correct, if the relevant articles found by citation searching also appear in your search you are on the right track. If not, you may want to consider adjusting your search to include these articles.
At a later stage, you can use a key-article to search the references (back in time) or citations (forward in time) to find articles that are relevant but not yet found by your search.
Many tools for systematic reviews have been develepod. An overview of many of such tools can be found in the SR Toolbox. Please note that not all these tools have been thoroughly tested or kept-up-to-date/maintained.
Below a few examples from the toolbox:
Systematic Review Accelerator; Using the Polyglot Search you can "translate" your searchstrategy from one database to another. A (free) account is required.
Medline transpose -- translates your PubMed- search to a MEDLINE-OVID search. It is also optional to translate the Boolean operator AND to 'ADJ' (a proximity operator).
EPPI Reviewer is a program
kmine literature explorer allows researchers in pharma, food and personal care industry to easily perform fast Scoping and Evidence in available literature.
A team of researchers from Utrecht University, in collaboration with the University Library and Information and Technology Services, is developing the ASReview tool. The aim of this tool is to speed up the screening of literature during a Systematic Review using Active Learning. Want to use AS review within the UMCU environment: use the Digital research environment (DRE).
Interested? Click here.
To pre-register your systematic review, you can use PROSPERO. You'll find ongoing reviews listed here as well. Since a full systematic reviews takes months to complete and publish, you can see if anyone is working on a similar topic, and prevent double work.
For a preferred reporting format, see the PRISMA guidelines. Some journals require that you follow the PRISMA guidelines, if not, they still help you work systematically. The final goal in reporting on a systematic search is being transparant and reproducible.
Covidence and Rayyan are two common programmes used for screening results of a systematic search.
Covidence is recommended by the Cochrane Community. The UU and/or UMC Utrecht does not have a license for the version that allows more options. The free version of Covidence allows you to upload 500 references and 1 review for 2 reviewers.
Rayyan is widely used, free and allows the upload of an unlimeted amount of references. One person (who coordinates the screening proces) needs to make an account and can then proceed to invite the other team members. The account verification is not automatic, so it can take minutes or in rare cases hours before your account is verified. Rayyan works fully online, through the website or the app. When there is a lot of traffic on the website this can lead to some delays. These issues are usually solved quickly.
The library has subscriptions to many journals, but not all. If you can't find full text via the UBU-link button, the PDF below has some handy tools to find them.
Chek for instance the FireFox/Chrome add-on: Google Scholar Button
