searching the literature systematically · searching the literature systematically taking a...

3
Searching the literature systematically Taking a systematic approach to a review involves including in your report/paper, your aim, your method of finding the papers from which you chose the ones you reviewed, the number of papers you found originally and how that changed as you went through your selection process. Knowing what you have to present can help you know what information to retain as you go through the search process. Search workflow Put your topic into the form of a research question. Identify the main concepts and their synonyms. These will be your search terms. Know what makes a paper one you will choose to use (inclusion criteria) and what makes it one you will not ( exclusion criteria). It may help to map your topic to “PICOS”, ie Patient Population or Problem, Intervention, Comparison, Outcomes, Study Design Use abstracting and indexing (A&I) databases’ Search histories to see what numbers and relevancy you get when you try different search terms and their combinations. Your initial (or Scoping) searches can change your research question because you may: i) Not see enough papers to be sure you can do a review on your topic. In which case, decide what interests you most and what concept you can drop or loosen or broaden. ii) See too many relevant papers. If so, consider what additional concept to add to get enough papers to review? When you have a search(es) which works for your research question, run it on the A&I databases you have chosen. Not everything has to go in as an A&I database search term, eg population age and methodology may be better assessed by reading the records or a whole article, rather than by relying on search terms. Use the A&I databases to get to a set of records which will include those for relevant articles and which is not too big for you to assess each record against relevancy and your inclusion/exclusion criteria. Assessment of even a few hundred records can take less time than you might expect. Use the Export/Download/Save function to collect the records in one place, eg reference management software such as EndNote online or Mendeley. RIS files are generic reference management software compatible files. These are the records you use for the rest of the process. Note the total number of records found. Note the A&I databases and the search terms used De-duplicate. You use more than one A&I database to get a breadth of results. In similar subject-related A&I databases there will be core journals indexed in all of them which means a search in each will return records describing the same articles. You don’t need to assess (screen) more than one record of the same article. Note the number of records left after duplicate records are discounted from your search results. Screening Note the number of records you have after you have read titles and abstracts to assess relevancy and inclusion and exclusion criteria. Note the number of records you have after you have read the full text of those remaining from your first assessment (based on titles/abstracts). Presentation of the numbers. Often the presentation of the numbers through the process is as a “PRISMA diagram” www.prisma-statement.org > Flow Diagram PRISMA diagram in an article: McInnes, E. et al (2012) Preventing pressure ulcers—Are pressure-redistributing support surfaces effective? A Cochrane systematic review and meta-analysis, International Journal of Nursing Studies, 49(3), 345-359. http://ezproxy.is.ed.ac.uk/login?url=https://doi.org/10.1016/j.ijnurstu.2011.10.014 See page 3 below

Upload: others

Post on 19-Jul-2020

7 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Searching the literature systematically · Searching the literature systematically Taking a systematic approach to a review involves including in your report/paper, your aim, your

Searching the literature systematically Taking a systematic approach to a review involves including in your report/paper, your aim, your method of finding the papers from which you chose the ones you reviewed, the number of papers you found originally and how that changed as you went through your selection process.

Knowing what you have to present can help you know what information to retain as you go through the search process.

Search workflow

Put your topic into the form of a research question.

Identify the main concepts and their synonyms. These will be your search terms.

Know what makes a paper one you will choose to use (inclusion criteria) and what makes it one you will not (exclusion criteria).

It may help to map your topic to “PICOS”, ie Patient Population or Problem, Intervention, Comparison, Outcomes, Study Design

Use abstracting and indexing (A&I) databases’ Search histories to see what numbers and relevancy you get when you try different search terms and their combinations.

Your initial (or Scoping) searches can change your research question because you may:

i) Not see enough papers to be sure you can do a review on your topic. In which case, decide what interests you most and what concept you can drop or loosen or broaden.

ii) See too many relevant papers. If so, consider what additional concept to add to get enough papers to review? When you have a search(es) which works for your research question, run it on the A&I databases you have chosen.

Not everything has to go in as an A&I database search term, eg population age and methodology may be better assessed by reading the records or a whole article, rather than by relying on search terms.

Use the A&I databases to get to a set of records which will include those for relevant articles and which is not too big for you to assess each record against relevancy and your inclusion/exclusion criteria. Assessment of even a few hundred records can take less time than you might expect.

Use the Export/Download/Save function to collect the records in one place, eg reference management software such as EndNote online or Mendeley. RIS files are generic reference management software compatible files.

These are the records you use for the rest of the process.

Note the total number of records found.

Note the A&I databases and the search terms used

De-duplicate.

You use more than one A&I database to get a breadth of results. In similar subject-related A&I databases there will be core journals indexed in all of them which means a search in each will return records describing the same articles. You don’t need to assess (screen) more than one record of the same article.

Note the number of records left after duplicate records are discounted from your search results.

Screening Note the number of records you have after you have read titles and abstracts to assess relevancy and inclusion and exclusion criteria.

Note the number of records you have after you have read the full text of those remaining from your first assessment (based on titles/abstracts).

Presentation of the numbers. Often the presentation of the numbers through the process is as a “PRISMA diagram” www.prisma-statement.org > Flow Diagram

PRISMA diagram in an article: McInnes, E. et al (2012) Preventing pressure ulcers—Are pressure-redistributing support surfaces effective? A Cochrane systematic review and meta-analysis, International Journal of Nursing Studies, 49(3), 345-359. http://ezproxy.is.ed.ac.uk/login?url=https://doi.org/10.1016/j.ijnurstu.2011.10.014

See page 3 below

Page 2: Searching the literature systematically · Searching the literature systematically Taking a systematic approach to a review involves including in your report/paper, your aim, your

A&I databases to use: www.ed.ac.uk/is/databases-subects > Nursing

Include: CINAHL Plus and MEDLINE (PubMed is the free interface to MEDLINE. You don’t need to use both). Include others which seem sensible.

Some will be on the same interface, eg MEDLINE, EMBASE, psycINFO and Global Health are on the Ovid interface.

Use Change to search them all at the same time. Use Deduplicate to have Ovid interface to create a set of records from an existing one which has had automatically identified duplicate records removed. If you export only the deduplicated set of results, remember to note the number of records found by your search across the multiple A&I databases for inclusion in your PRISMA, etc diagram.

A&I database search tips Map term to subject heading/Suggest subject terms

Links from the Library to many of the health databases take you into Advanced Search with a default setting which means the search you do will show you a list of the A&I database’s subject headings which match your search term. Not a list of records which contain your search term.

Looking at a particular A&I database’s subject headings is a good way of adding to your search terms.

Ovid MEDLINE(R) Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Ovid MEDLINE(R) Daily and Ovid MEDLINE(R) 1946 to Present

If there’s a subject heading used by the database for the concept(s) which you want to read about, adding them

to the search terms you’ve already thought of, will add more and relevant records. They are a good prompt for reminders of USA terminology.

“Take the tick off” if you’re happy you’ve got the search terms you want. Doing this and using a subject heading as the search term will get you all the records you would find going through the route with the “tick on”, plus records where no subject heading field has been added to a record yet but where the phrase appears in title or abstract etc. Use “phrase marks” if you want the words you enter to appear together as you’ve typed them. Use the truncation symbol, which is usually the asterisk, *, to look for words with variant endings, any of which is useful. You then don’t have to type them all in and you don’t have to remember them all either, eg geriatric* rather than geriatric/geriatrics/geriatrically… Use the Search histories to begin with. Look at each concept separately and then put them together. The advantages of doing this are:

You see which concept has the smallest number of records (which indicates why a search may have zero results).

You can add synonyms and extra concepts as you think of them.

You don’t have to worry about how the search interface priorities commands. You need to “bracket off” OR clauses in a one string search, eg (geriatric* OR elderly) AND (mobil* OR exercise) returns records which mention one of each concept. geriatric* OR elderly AND mobil* OR exercise will return records with the words elderly and mobil* but also any record with only one of the other words, ie geriatric* or exercise.

OR = more results (all the grey area )

AND = less results (just the grey area )

Hyphens and spaces are often treated the same. Do a search to check.

Page 3: Searching the literature systematically · Searching the literature systematically Taking a systematic approach to a review involves including in your report/paper, your aim, your

Inter-Library Loan (I.L.L.) service to get books or journal articles, etc, which the Library does not have all. Register online at https://ed-ac.illiad.oclc.org/illiad

Using the “First Time Users” link. You will be asked to select a delivery location which will be the library site (eg Main Library) from where you will collect original items.

You can also use this system, “ILLiad”, to request scans of articles from journals we only have in print in the Library Annexe [Free. No annual limit].

You get notified when your registration is successful and thereafter login to submit requests, track their progress and change your contact details on the system.

. Copies of articles etc will be sent to your chosen contact address.

. You can request they are posted to you.

. Scans of articles can also be delivered electronically.

Your first 30 I.L.L. requests in a year will be free [5 for undergraduates].

After that, each successfully supplied request will cost you a (subsidised) £5. For which (probably monthly) you will receive an advice note listing items which have been supplied to you during the previous month. Payment is via the University's e-pay system (www.epay.ed.ac.uk).

. ILLs usually take 2-3 working days to arrive.

. More information from the link at: www.ed.ac.uk/is/inter-library

Example of thinking about search terms:

What is the experience of adolescents with a history of disruptive behaviour undergoing a course of mindfulness training?

adolescents disruptive behaviour Mindfulness training experience

adolescen(ts)

teen(agers)

“young people”

youths

“disruptive behav(iour)”

aggress(ion)

violen(ce)

mindful(ness)

Meditation

Training/education/program(me)/course…

attitude(s) / satisfaction

Compliance

Dropout(s)

Help

Contact me if you would like further advice or guidance on your own systematic review.

[email protected], Academic Support Librarian (Health in Social Science) CINAHL Plus search tips: www.docs.is.ed.ac.uk/docs/Libraries/PDF/NursingMScCINAHL.pdf

Introduction to systematic literature searching is an in-house guide: www.docs.is.ed.ac.uk/mvm/IntroSystematicLitSearch.pdf

EndNote desktop is on the open access computers. In the Start menu. It is also freely available to download for your own devices. www.ed.ac.uk/information-services/computing/desktop-personal/software/main-software-deals/endnote

Close all Microsoft Office applications before installing. Compatibilities - https://endnote.com/product-details/compatibility It has more functions than EndNote online. See a comparison table at: www.docs.is.ed.ac.uk/mvm/BiblioManagersTable.pdf EndNote online is freely available and may be useful for backing up and/or sharing a desktop library. www.docs.is.ed.ac.uk/docs/Libraries/PDF/guideEndnotewebregistering.pdf

Sep19

The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.