Developing a search strategy
Cari
and Lee look at articles in the gold set carefully, and identify any alternative terms relating to their topic
of interest. They arrange the terms into the PICO table Cari started to develop earlier.
|
|
MeSH terms & Subject headings
Once Cari and Lee have their main terms, they use the database thesauri (subject headings/Medical subject headings/MeSH) to see if there are any other terms they can add.
When using Ovid MEDLINE, Cari searches for each of her keywords using the 'Map Term to Subject Heading' option.
 |
Cari is not sure that all of the given subject heading suggestions are applicable or relevant. In these cases, she uses the scope notes within the database to learn more about the heading and what it relates to. One example of this was while looking for ‘cholesterol’ subject headings. Cari was offered the option of ‘Cholesterol esters’.
From this, she learns:
Cari discusses this with Lee, and they decide to include the term in their search strategy.
Cari finds that by clicking on a particular subject heading link she will be taken to a subject heading “tree” that illustrates the relationship between subject headings. She finds that Cholesterol Esters is a narrower term of Cholesterol, and that she can include all of these Cholesterol terms in her search by “exploding” the main subject heading, Cholesterol:
Cari adds any relevant subject headings she finds to her PICO table. She indicates which terms are subject headings by writing a slash / after them. She also adds in any new keywords that she has found, and identifies where any spelling differences might arise, or where truncation might be needed to find all relevant articles. For example using theobrom* will extend her search to locate papers containing either the keyword theobroma OR theobromine.
Attempting the first search
Now Cari attempts her first search:
| 1st step | She enters each search term separately to record how many results are retrieved for each one - this also assists her to avoid making any typing errors. |
| 2nd step | Once all terms for a particular concept are included, she combines them with OR. |
| 3rd step |
When all the concepts have been entered in this way, she combines each concept with AND. |
As this search took a while to organise, Cari creates a free account with Ovid, so she is able to save her search. She also saves a PDF copy of the entire search in her electronic lab notebook.
Some reflections
Lee and Cari now check that the search has retrieved relevant results, and also picked up all the appropriate articles from their gold set.
Wanting to make sure she hasn’t missed anything obvious, Cari obtains a search strategy appraisal checklist and works through it to see if there is anything else in her search that needs amending.
As she looked through the results, Cari realised that using the search term ‘flavanols’ has brought up articles referring to various berries and tea, as well as cocoa.
Cari and Lee discuss the idea, and decide to investigate proximity operators, once they have done more searching. Looking for certain word combinations close to each other might help them to filter out irrelevant articles. Otherwise, they can do this manually during the screening phases of their review based on their inclusion and exclusion criteria.
Eventually, while looking through their initial search results, Cari and Lee realise that there are a lot of articles about flavonoids and polyphenols that don’t discuss cocoa at all. They test the search taking out polyphenols and polyphenols/, ‘flavanols’ and flavonoids/, and find that they still retrieve the relevant articles from their gold set, but find a lot less irrelevant articles.
They decide to omit these terms from their search strategy.
Filtering and limiting
Now that Cari and Lee are happy with their search terms, they look at the results to see how they need to filter them. Cari had already decided that articles would have to be available as English language full text, as she doesn’t have access to a translator.
Cari and Lee also need to determine their list of databases to search. As there are many databases, each with specific emphases, Cari and Lee decide to meet with a specialist librarian to get expert advice. They are also interested in any areas where grey literature might be of use to them, such as reports from organisations concerned with cardiometabolic health, obesity or cholesterol. Cari and Lee are also keen to ensure they collect any relevant records of theses, dissertations or conference proceedings. The librarian also gives them some tips for setting up alerts in the databases, so they’ll hear about any new articles that are published after they complete the main search phase.
