Overview
This page is all about saving you time while improving the quality of your searches.
Recommended database search workflow
The diagram below represents one approach to systematic searching. It incorporates factors to consider, along with UniSA Librarians' tips to help make the process more efficient and your search more comprehensive, transparent, and reproducible.
Easily edit searches with Ovid Search Launcher
Ovid Search Launcher allows you to enter your search in plain text, and run it in any UniSA-subscribed database on Ovid.
The Launcher provides a quick and easy way to edit your searches - move the position of search lines, add and remove terms, etc.
Avoid inconvenient timeouts while editing on the Ovid platform, especially if working with longer, more complex strategies.
Simply paste a plain text version of your search strategy or search history into the search launcher, and from the 'Choose databases' drop down menu select the database you want to search, and then click outside the menu to retain your selection. The codes below align with the UniSA subscriptions. It is recommended to search one database at a time.
- MEDLINE - medall
- Embase - emczd
- Emcare - emcr
- PsycInfo - psyh
- JBI - jbi
-
Ovid Search Launcher
Use this link to access the launcher outside the UniSA network. The version of the launcher freely available online will not work offsite (even if you select Open Athens and UniSA)
Top tools
You may find the tools below useful through the search process.
-
PubMed PubReMinerA tool that allows simple text-based query building and information gathering (mining) of the NCBI literature search engine PubMed.
-
MeSH BrowserUse the free MeSH Browser to see where terms appear within the theasurus.
-
Smart quotes in WordInstructions from Microsoft on how to disable 'smart quotes' in Word - straight quotes are needed when copying and pasting into search interfaces
-
Plain text editorLook on your device for a plain text editor such as Notepad (PC), TextEdit (Mac). These tools allow for easy editing and reuse of search strategies
-
RobotSearchA validated tool (machine learning model) that identifies reports of randomized controlled trials
Validation / sample set of studies
A sample set of references that would be eligible for inclusion in your review can be a valuable tool. You can:
- mine the database records for these references for relevant 'free text' terms (aka 'keywords') and subject headings
- test your search performance by seeing if your draft search strategy would locate your sample papers
Developing your sample set:
- consult your review group and subject experts
- preliminary searching
- reading
- other reviews
'Mining' your sample set - search for the reference in MEDLINE and see if it is included (tip: search by Title). If it is, look at the Complete Reference and see what terms appear in the Title, Abstract, Subject Headings, Keyword Heading and any other relevant fields.
Testing search performance - search for the reference in MEDLINE and see if it is included (tip: search by Title). If it is, combine it with the final line of your search strategy using AND. If the record still appears, your search would have located this record, which is a positive indicator. If not, then review your strategy to see why this might have occurred (e.g. a missing keyword or subject heading, or too narrow a search).
You may not be able to edit your strategy to locate all of your test set - for example, some records are poorly indexed. However you should be able to locate most.