Scoping searches are fairly brief searches of existing literature designed to help you gain an overview of the range and depth of research that exists for a particular research idea. It can cover published work and discover on-going studies. Research proposals are shaped by the results of a scoping search for your research idea. In short, it should ensure that your research complements and adds to existing research so that your proposal is not rejected for ‘re-inventing the wheel’. Check to make sure no one else has undertaken a systematic review on your topic before, and check that there is enough evidence for a meaningful review. Places to check include the Prospective Register of Systematic Reviews PROSPERO, the OSF and the Cochrane Library.
Systematic reviews have clearly defined questions, with an inclusion and exclusion criteria. Before you start your search you will decide what types of studies you will include, for example, RCTs or qualitative.
There are frameworks to help you formulate your question:
PICO = Patient/Population Intervention Comparison Outcome, you sometimes see PICO as PICODS where D is the study design and S is the setting.
ECLIPSE = Expectation, Client group, Location, Impact (what’s changed since the policy), Professionals, SErvice. Developed for health policy and management.
PEO = Population, Exposure, Outcome. Developed for qualitative searches.
PESTLE = Political, Economic, Social, Technological, Legal, Environmental. Developed for management and strategy.
SPIDER = Sample, Phenomenon of Interest, Design, Evaluation, Research type. Developed for qualitative searches.
A protocol is developed at the start of the process as a plan of the work to be undertaken, decisions are made about the inclusion and exclusion criteria, the databases that will be searched etc. The idea of the protocol is to reduce bias as well as well as notify other researchers of your intention to undertake this work, hence the requirement to register the protocol in PROSPERO or OSF.
See, Lasserson, T.J., Thomas, J. and Higgins, J.P.T. (2019) 'Starting a review', in Cochrane Handbook for Systematic Reviews of Interventions, pp. 1-12. Available at: https://doi.org/10.1002/9781119536604.ch1.
As well as the PRISMA for systematic review protocols (PRISMA-P) guidance for the development and reporting of systematic review protocols at https://www.prisma-statement.org/protocols
Your supervisor should be able give further guidance for your protocol if needed.
You will search bibliographic databases (Web of Science, APA PsycInfo, CINAHL, Medline etc.) to locate journal articles and conference abstracts on your review topic. You will create a search strategy that will search the 'free-text' database fields that represent the article, such as Title, Abstract, Keywords, which is not usually the whole article, as well as the thesaurus terms, sometimes referred to as subject headings or descriptors. Your search will be included in an appendices so that it is transparent and reproducible.
If you look at a Cochrane Library Review you will see the search strategies for the different bibliographic databases listed. The search commands, or syntax, vary depending on the platform you are searching so the search commands you see for Medline (Ovid) database will be slightly different to Medline (EBSCO) database.
The Library subscribes to many databases, check what is on offer via our subject guides where databases are grouped by subject.
To reduce bias consider including databases from outside the Global North, such as...
Use Boolean operators OR, AND, NOT to combine your search terms together.
Please watch this short Video for further explanation.
You often see NOT used to remove animal studies from the search results (more to follow on this in the Search Filter section below).
A good systematic review will include both a free-text search as well as an thesaurus (subject heading/descriptor) search. When you are searching free-text, you are searching the database fields that represent the whole article, usually the Title, the Abstract and the Keywords.
The building block of a search strategy are your keywords or the concepts, these can be "Natural" or "Academic" language; a good example of this is teenager OR adolescent. A thorough systematic review will have a comprehensive free-text search at its heart where all the appropriate synonyms of a keyword have been used. Also, consider UK and American spellings.
These are symbols such as * $ ? that are used to replace a single letter, a missing letter or a series of letters in a word. When the symbol is placed in the word it is a wildcard and when it is at the end of the word it is a truncation symbol. Wild cards can be useful for capturing the UK and American spellings.
A wildcard is useful to:
Using a truncation symbol, usually a * but sometimes a $, will search for the main root of the word and any variable endings. For example…
Assess* will find articles containing assess, assessing, assessed, assessment.
The symbols vary according to the database you are using so it is important to check the Help section in each database before you start.
"Proximity" or "adjacency" can be a very useful tool when you are running your free-text searches. It is a way to search for two or more words that occur within a certain number of words from each other. This can be very helpful when your topic can be described in different ways. For example, 'climate change' could also appear in a paper as 'change in the climate'.
The Web of Science database uses the command NEAR/x where x is the maximum number of words that separate the terms from each other so to search for 'climate change' or 'change in the climate'. You would search... climate NEAR/3 change
Other databases use different terms for this.
The EBSCO platform uses Nx, x being the number of words in between, so... neonat* N3 care will find, neonatal care as well as, care of the neonate
The Ovid platform uses adjx, x being the number of words each term is separated by, so... neonat* adj3 care, will find neonatal care, as well as, care of the neonate
The University of Alberta Library has produced this excellent Searching syntax guide for common database platforms.
Many bibliograhic databases have a thesaurus and use this controlled vocabularly to describe the content, these subject headings or descriptors are assigned to the database record for the article retrospectively, after the full article has been scanned and the subject headings are then assigned to describe it using consistent language. Each database will have its own subject specific thesaurus, for example the Medline database (PubMed) uses Medical Subject Headings, you will hear these referred to as MeSH, these subject headings are from a controlled vocabulary decided by the Nationa Library of Medicine.
You will need to check the subject headings (descriptors) for each database as they will vary, for example, the subject heading for allergy in the Medline database is Hypersensitivity/ but the subject heading for allergy in the EMBASE database is allergy/
Search filters, sometimes called a search hedge, are predefined search strategies that have been created by information professionals. The searches will vary depending on the database and platform. Some patforms, Ovid is a good example, provide these as pre-saved searches that you can run and include in your search using AND, see Expert searches at https://tools-ovid-com.soton.idm.oclc.org/ovidtools/expertsearches.html
For others, there is an excellent resource called the ISSG Search Filters Resource https://sites.google.com/a/york.ac.uk/issg-search-filters-resource/home. You will filnd filters for AI, Geography, Gender to name a few.
You may need the search to exclude animal studies presented using the NOT operator, this is beause using the limit to/refine to 'human' function in the database may lose some relevant studies. The first part of the search is to command the database to search for animal studies but not exclude the human studies and then NOT this with your final results.
For the Medline database via Ovid platform the command is... exp animals/ not humans.sh.
So your search will look something like this...
9 7 and 8
10 exp animals/ not humans.sh.
11 9 not 10
Please see section 4.4.b of the Cochrane Handbook https://training.cochrane.org/handbook/current/chapter-04#section-4-4
For the Embase database via the Ovid platform the command is... not ((exp animal/ or exp invertebrate/ or nonhuman/ or animal experiment/ or animal tissue/ or animal model/ or exp plant/ or exp fungus/) not (exp human/ or human tissue/))
So your search will look something like this...
9 7 and 8
10 ((exp animal/ or exp invertebrate/ or nonhuman/ or animal experiment/ or animal tissue/ or animal model/ or exp plant/ or exp fungus/) not (exp human/ or human tissue/))
11 9 not 10
Grey literature can include documents created by government departments, professional bodies, charities, non-governmental organisations, royal colleges, businesses, and industries. Examples include interviews, research reports, blogs, working papers, podcasts, conference abstracts, theses, white papers, policy documents, guidelines, clinical trials, and social media posts. See our separate grey literature page for further information.
You may hear citation searching, referred to as citation chaining, citation chasing, snowballing or pearl growing! You might look at the references of your included articles, which is backward searching, and you might see who has cited your articles since publication which is forward searching. The database record might contain a link to the citing articles in that database. The Web of Science and Scopus databases are good for tracing citing articles, and the results can easily be exported into your reference manager and/or screening software.
There are online tools you might explore too:
The Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) statement is designed to help you make your methodology clear and complete. Please see https://www.prisma-statement.org/prisma-2020. PRISMA-P has been developed for guidance with the protocol and the PRISMA flow diagram will help you demonstrate how you screened your search results down to your included studies.
