Systematic Reviews: Drawing up your search strategy

• Use of both free text terms and thesaurus terms

• Exploding thesaurus terms to capture narrower terms

• Generous use of adjacency operator

• Aim to maximise sensitivity of search strategy

• Uses multiple databases to ensure maximum number of relevant studies are identified

"Once the concepts of the search have been determined, the next stage is to produce a list of synonyms, abbreviations and spelling variants which may be used by authors. Similar research is often described using very different terms. To reflect this variation, a search strategy will usually comprise both indexing terms (if the database has a thesaurus or controlled vocabulary) and ‘free text’ terms and synonyms (from the database record’s title and abstract) to ensure that as many relevant papers are retrieved as possible. For example, when searching MEDLINE for studies about myocardial infarction, the free text term “heart attack” should be used as well as the Medical Subject Heading (MeSH) term “Myocardial Infarction”. Identifying appropriate indexing terms can be done by searching for key papers and checking how they have been indexed, consulting clinical experts in the review team and advisory group, as well as by scanning the thesaurus for relevant terms."

See the example search strategy  from the Centre for Reviews & Dissemination  which is available on this web page via the PDF called 'Guidance PDF'. See Appendix 2 which provides a detailed discussion of how to structure a search from a review question, including the use of search filters for study design.

NHS Centre for Reviews & Dissemination. Systematic Reviews: CRD's Guidance for Undertaking Reviews in Healthcare. York: University of York NHS Centre for Reviews & Dissemination, 2009.

See also How to develop a search strategy for an intervention review

   - From Cochrane Effective Practice and Organisation of Care  (EPOC) group resources for authors https://epoc.cochrane.org/resources/epoc-resources-review-authors

The goal of systematic review searches is to identify all relevant studies on a topic. Systematic review searches are therefore typically quite extensive. However, it may be necessary to strike a balance between the sensitivity and precision of your search.

See also:  IOM Standards for Systematic Reviews: Standard 3.1: Conduct a comprehensive systematic search for evidence

 

Antidepressant drugs OR antidepressive agents	Eating disorders AND cognitive therapy

AND and OR are Boolean operators.

OR is used to combine synonyms, abbreviations and all related terms on a similar concept. You can OR together subject headings for a particular concept with relevant keyword searches. 'OR is more'. Your result set will get larger as you OR together more terms.

AND is used (normally at the end of the search) to combine together different concepts and to retrieve results where all the concepts are present. AND narrows down your results and makes your search more specific. AND is sometimes automatic for two or more search terms depending on the database.

For more information on boolean operators see the  literature review library guide

Note: "The NOT operator should be avoided where possible to avoid the danger of inadvertently removing from the search set records that are relevant. For example, when searching for records indexed as female, 'NOT male' would remove any record that was about both males and females."

From: Cochrane Handbook for systematic reviews of interventions Section 6.4.7 Boolean operators

• A systematic approach to searching: an efficient and complete method to develop literature searches

  Creating comprehensive search strategies for different databases.

• Pearl growing  Citation pearl growing is the process of using the characteristics of a relevant and authoritative article, called a pearl, to search for other relevant and authoritative materials
• Snowballing – scanning reference lists and citation tracking
• Using dictionaries and thesauri

• Fixed list of terms arranged hierarchically with broader and narrower terms.

• Indexers classify the article by tagging it with subject headings that relate to the content.
• Some tags represent the main focus of the article and some refer to secondary aspects of the work.

• Can allow you to search more effectively and avoid missing relevant articles.

• Can retrieve relevant articles where the term does not occur in title or abstract.

• Include synonyms and abbreviations

• Use British and American spelling and terminology

• Use singular and plural forms

• Use truncation wildcard searching

• Use software mapping features e.g. MeSH in PubMed

• Look at scope notes for each thesaurus term and note related terms

• Explore hierarchical thesaurus trees to identify most helpful level

• Look at how other papers have been indexed

Cochrane recommends "In order to identify as many relevant records as possible searches should comprise a combination of subject terms selected from the controlled vocabulary or thesaurus (‘exploded' where appropriate) with a wide range of free-text terms."

From Cochrane Handbook for systematic reviews of interventions 6.4.5 Controlled vocabulary and text words

• article title
• abstract
• keywords

Cochrane systematic reviews and those systematic reviews considering interventions often include a filter to restrict the results to studies reporting Randomised Controlled Trials (RCTs). Consider whether you wish to include a filter for a particular type of study design as part of your search strategy. These exist for RCTs but also for other study types including Observational Studies and Patient Issues. For more information about where to source standardised, pre-tested search filter search strategies optimised for different databases which you can copy and paste and add to your own search strategy see the Using Filters tab of this guide.

Explode: 

• Select Explode to automatically include all articles which have also been tagged with any more specific narrower terms in the thesaurus under the subject heading you have chosen to search.
  • The indexers will select the most specific subject heading possible to tag an article so by choosing explode you can select a high level subject heading (e.g. antibiotic agent is used in Embase for the concept antibiotic) and by exploding the subject heading you will automatically include the narrower, more specific terms, e.g. named antibiotics.
• To view the more specific subject headings which would be included click on the subject heading itself to view it in the thesaurus. The structure of the thesaurus may be displayed in different ways depending on the database, e.g. in Embase on the Ovid platform these are listed under Narrower Terms; in Medline narrower terms are indented in.
• In general it is good practice to explode the Subject Headings in your search. If you feel that too much irrelevant material is being retrieved then explore the thesaurus to see whether you need to pick a high level subject heading (not exploded) and also only some of the more specific subject headings which fall beneath it in the thesaurus.

Focus (in the Ovid databases):

• If you select Focus you will restrict your results to only those articles which the indexer feels the subject heading you have selected to search is key to what the article is about. 
• In other databases Focus may be known as something else, e.g. Major Concept in CINAHL on the Ebsco platform.
• On the Ovid platform if you view the Complete Reference for an article you will see all the subject headings which have been assigned to an article, and the ones which are key (which would be retrieved by a focused search) are marked by an asterisk before the subject heading. In CINAHL if you view the Detailed Record you will see them listed under Major Subjects.
• Use Focus with care in a systematic review as it will dramatically reduce the number of results retrieved. Initially it would be a good idea to see what the results are without focusing after you have combined your terms. You may find it useful to use Focus for very peripheral aspects of your topic which you wish to include without being inundated with results.

It is possible to both Explode and Focus a subject heading search.

• By applying both Explode and Focus you will retrieve articles tagged with your subject heading and the narrower subject heading terms that fall underneath it in the thesaurus (Explode), but also limit (reduce) the results to where either the top level subject heading or any of its narrower terms have been identified by the indexers as being key/the focus/the major concept which the article is about.

A search on Embase on the Ovid platform showing the initial Subject Heading selection screen. Explode and Focus are available to select:

By clicking on a specific Subject Heading in the list, e.g. antibiotic agent, you enter the thesaurus and can see what the term is used for, the broader terms and narrower terms. By selecting Explode for antibiotic agent in Embase you will automatically retrieve articles which have also been tagged with all of the narrower more specific named antibiotics.

Below you can see the Full Record for a particular article (in Embase on the Ovid platform) with the Subject Headings listed which have been used to tag this particular article. You can see that 'penicillin allergy' is one of the subject headings assigned to this article and the asterisk * before the subject heading indicates that this was deemed to be key to what the article was about when it was added to the database and subject headings assigned. A normal subject heading search for 'penicillin allergy' or its exploded broader term would retrieve this article but so would a focussed search *penicillin allergy/. An exploded and focused search for the broader term, e.g. drug hypersensitivity, would also retrieve this article - exp *drug hypersensitivity/.  

It is always worth checking to see whether any systematic reviews which have a concept in common with your search question have published their search strategy. If they have then this will act as a useful starting point for you to use for your search. Remember it is not necessary for all the concepts of the systematic review to be the same as you should be able to isolate the specific lines of the search strategy relevant to you. So for example if I am undertaking a systematic review on the 'effectiveness of phototherapy for neonatal jaundice' I may find that part of a search strategy for a published review on 'Phototherapy for treating pressure ulcers' is very relevant for me. I will then need to either create a search strategy element for my other concept of 'neonatal jaundice' or locate another systematic review which may have already created a search strategy including an element for this concept, e.g. one on 'Early intravenous nutrition for the prevention of neonatal jaundice'.

Below is an excerpt from a published Cochrane systematic review on 'Phototherapy for treating pressure ulcers' showing the search lines for the phototherapy concept (optimised for Ovid Medline) which could then be used in a different search and combined with other concepts.

Chen C, Hou WH, Chan ESY, Yeh ML, Lo HLD. Phototherapy for treating pressure ulcers. Cochrane Database of Systematic Reviews 2014, Issue 7. Art. No.: CD009224. DOI: 10.1002/14651858.CD009224.pub2.

Systematic reviews in the Cochrane Library should all publish their search strategies and in many cases will show the search strategy they used for each database they searched (optimised for each database) - this normally appears in the Appendix of the full text for a Cochrane review.

The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) checklist states that Systematic Reviews should "present full electronic search strategy for at least one database, including any limits used, such that it could be repeated". However, unfortunately you may find that older systematic reviews and even some current ones may not do this. Note too that in some cases the full search strategy may appear as a supporting document on a journal site as opposed to forming part of the pdf of the article. 

Pay attention when looking to use a published search strategy to both the database and the platform it is hosted on and for which the search strategy has been optimised. If you are searching the same database on the same platform then you should simply be able to copy and paste the search strategy in to generate results. If you are uncertain about what the lines of the search strategy mean and what they are searching, e.g. exp, adj or .tw then see the information on the Advanced search techniques tab.

Remember when using a search strategy from another systematic review that you should assess it for quality rather than simply copying it in. Are there any other search terms that should be included, have they located all appropriate subject headings, and so on? See the information about PRESS: Peer Review of Electronic Search Strategies on the Advanced search techniques tab. The age of the source systematic review is also important as subject headings change to reflect changes in medical science. As such if you are working in an area of rapid development and the strategy is a few years old you may wish to just use it as a starting point rather than directly copying.

Finally if you do use all or part of another's search strategy then do remember to acknowledge the source.

No guidelines exist on when to stop searching. The following articles might give some assistance in deciding

when "enough is enough":

How much searching is enough? Comprehensive versus optimal retrieval for technology assessments.

Booth A1. Int J Technol Assess Health Care. 2010 Oct;26(4):431-5. doi: 10.1017/S0266462310000966. Epub 2010 Oct 6.

Estimating the Horizon of articles to decide when to stop searching in systematic reviews: an example using a systematic review of RCTs evaluating osteoporosis clinical decision support tools.

Kastner M, Straus S, Goldsmith CH. AMIA Annu Symp Proc. 2007 Oct 11:389-93.

The capture-mark-recapture technique can be used as a stopping rule when searching in systematic reviews.

Kastner M, Straus SE, McKibbon KA, Goldsmith CH. J Clin Epidemiol. 2009 Feb;62(2):149-57. doi: 10.1016/j.jclinepi.2008.06.001. Epub 2008 Aug 22.