Guest blog: How to get started with literature searching
Josh, who wrote this guest blog taking some time out from literature searching to do some mountain hiking!
by Josh Cheyne, Information Specialist
“I write search methods for systematic reviews (SR) and support the publication of Cochrane SRs written by authors all over the world. I started working on a freelance basis in addition to my Cochrane commitments last year. In my previous job as an NHS librarian, user education/ one-to-one sessions took up about 40% of my time. I still find search methods a difficult concept to describe - there’s so much which could be covered. Every enquiry is unique and cannot be generalised or addressed with a one-size-fits-all approach.
At the beginning of your research, get help from a skilled Information Specialist or Librarian. They will advise on constructing high quality search methods and search strategies and how to best managing the results of the searches. We work within the NHS, universities, colleges, trade unions, and on a freelance basis.
For this post, I’ll briefly describe some of the most frequent questions and problems I encounter when collaborating with researchers in healthcare and medicine.
Creating a clear question
Having a clear and defined question is central to the success of your topic. Vague questions, or questions that are too broad or general, can’t be answered easily. Worse still, when you start searching, you’ll likely retrieve an overwhelming number of results!
P.I.C.O (acronym meaning: Patient/Population/Problem.Intervention.Comparison.Outcome) is one of the most common frameworks to help define a clinical question into specific components and help develop a search strategy.
Here’s an example of a query:
“Left atrial appendage closure for prevention of stroke in atrial fibrillation (AFib)”
Population of interest/ Patient or the Problem: patients at risk of stroke (primary/ secondary)
Intervention: left atrial appendage closure (also known as LAA closure or LAAC)
Comparison: unclear/ current surgical or pharmacological protocol
Outcome: stroke/ recurrent stroke
Scoping/ reconnaissance searches
Reconnaissance searches are an essential part of the search process. This is where you see what’s there and the kind of evidence you’re going to need.
Take one letter from your PICO at a time. Start with what you know and make a list of important terms, keywords (the words you’d use to search Google), and synonyms for each component of the question.
Some of the terms you’ve chosen will be the most important terms used for describing each concept. Some terms may not retrieve very much, but don’t rule anything out just yet.
Try some simple searches using these terms and see what you retrieve. While you may start with Google, the following databases are a great place to start and the resources you search depend on the kind of question you’re trying to answer:
PubMed: the US National Library of Medicine’s database. It contains over 25 million references to journal articles in biomedicine, healthcare, and life sciences.
TRIP (Turning Research Into Practice) a free clinical search engine with basic and advanced search functions which searches multiple evidence types, including guidelines, evidence-base synopses, and primary research.
NICE Evidence search NICE guidelines and pathways; Sources include: British National Formulary (BNF), Clinical Knowledge Summaries, SIGN guidelines, the Cochrane Library, Royal Colleges, Social Care Online, and GOV.UK.
The Cochrane Library: Cochrane Database of Systematic Reviews (CDSR) includes all Cochrane Reviews (and protocols) prepared by Cochrane Review Groups. Cochrane Central Register of Controlled Trials (CENTRAL) is a repository of reports of randomized and quasi-randomized controlled trials populated by centralised searching services, searches conducted for systematic reviews, and searches conducted for Cochrane group specialised registers)
PROSPERO (International prospective register of systematic reviews) - this is a database of prospectively registered SR protocols which are recorded and maintained as a permanent record. If you are planning a SR, check here to see if a similar one already is in progress.
Starting a search
I tend to search within the title field only using a few important phrases or terms to do with the topic. If you’re not finding what you expect, it could be a couple of things:
· You might be looking in the wrong place: try a different database.
· You might be using the wrong terms to search for it: try alternative terms.
· or sometimes the research just isn’t there. A useful finding in itself!
Focus on getting the right words which best represent your question – worry about the numbers later on, otherwise you’re sacrificing quality for convenience. It sounds counter intuitive, but I find it helpful not to pay too much attention to the number of documents I retrieve during this stage.
Keep a detailed log of your searching activities throughout each stage of the process: where you’ve looked, the keywords you used, what you found, and how much.
You won’t find everything you need at this stage. Besides, this is just the start!
BOOLEAN, keywords and indexing: what’s the difference between keywords and indexing terms?
At the most fundamental level, database searches are constructed from keywords and Boolean operators: these are simple words (AND, OR, NOT or AND NOT) used as conjunctions to combine or exclude keywords in a search,
You can build very simple searches with only a few terms or you can construct much larger and more complicated search lines and larger strategies. The complexity of your search depends entirely on the question you’re trying to answer.
Keywords are the words present in an article’s fields within database records, such as title, abstract, or author.
Controlled vocabulary/indexing terms are selected lists of standardised terms and phrases used to index the contents of databases. They are a way of retrieving articles that may use different words to describe the same concept[AL1] .
PubMed uses Medical Subject Headings (MeSH), the National Library of Medicine's (NLM) controlled vocabulary. It provides a single entry point (controlled term) to search for medical concepts that appear in the literature under a variety of different terms or concepts.
The MeSH database also offers both definitions (Scope Notes) and a listing of synonyms and related terms for each MeSH term (‘Entry terms’), which lists different terms being used for a concept. Paracetamol is a great example to use here – the MeSH being "Acetaminophen"[Mesh], but obviously this retrieves records about paracetamol, too.
1. NLM’s MeSH: https://www.nlm.nih.gov/mesh or https://www.ncbi.nlm.nih.gov/mesh/
2. MeSH Browser: https://meshb.nlm.nih.gov/search
Where available, search using controlled vocabulary and keywords together. Controlled vocabularies are not available in all bibliographic databases. Here are a couple of examples using the question from the above P.I.C.O. example using MeSH headings and keywords.
AND links search terms and. Use this to narrow your search; ("Stroke"[Mesh]) AND "Atrial Fibrillation"[Mesh:NoExp]
OR finds multiple different terms (synonyms, acronyms). Use this to expand your search: synonyms, alternative terms, different spellings; (cerebrovascular*[Title/Abstract]) OR (stroke*[Title/Abstract])
NOT eliminates records which contain terms you don’t want (use with caution); ("Stroke"[Mesh]) NOT "Myocardial Infarction"[Mesh]
You can truncate terms to search for words beginning with the same word stem, eg. thrombo* will locate thrombosis, thrombolytic, thromboembolism, thrombocyte
Wildcard functions help search for alternative spellings (American and UK English) by replacing a character with a code (?, #, *), eg. ischaemic and ischemic paediatric and pediatric, or oedema and edema. The character and its function (whether it replaces zero, one, or more than one character) used will vary in each database. Always refer to the help section of a database for the specific truncation symbols and search syntax.
Comprehensiveness can be appropriate or not depending on the type of study you’re doing. For Cochrane systematic reviews, the searches have to be as comprehensive and exhaustive as possible, but if you were wanted to answer a specific clinical question for a patient, you wouldn’t need to be as comprehensive and retrieve absolutely everything. You can search for everything. It’s possible, but undesirable and impractical in the extreme. Remember, you’re the one who’s going to look through the results!
You can do it!
Searching for the most useful information can be a daunting process, but it is an iterative process, too, and requires some trial and error. Finding the appropriate balance between sensitivity and specificity can be very difficult: did you search enough databases, use the right words, use a Boolean operator in the right place? The important thing is to keep your methods consistent, document your search process from start to finish, and ask for help from an expert if you need it.”
Core Reading:
Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.2 (updated February 2021). Cochrane, 2021. Available from www.training.cochrane.org/handbook.
PRISMA Statement - Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097. doi: https://doi.org/10.1371/journal.pmed1000097.
Methodological Expectations of Cochrane Intervention Reviews (MECIR) (last updated February 2021). https://community.cochrane.org/mecir-manual
Bramer WM, de Jonge GB, Rethlefsen ML, Mast F, Kleijnen J. A systematic approach to searching: an efficient and complete method to develop literature searches. J Med Libr Assoc. 2018 Oct;106(4):531-541. doi: https://doi.org/10.5195/jmla.2018.283. Epub 2018 Oct 1. PMID: 30271302; PMCID: PMC6148622.