Skip to main content
 

Systematic Reviews: Using filters

Search Filters

People doing systematic reviews often use filters to only include randomized controlled trials which are widely seen as the best kind of research for evidence-based medicine. However, search filters exist for many other types of study design including Observational Studies and Diagnostic Studies.

The Cochrane Library Handbook lists filters for RCTs in the Medline database on the Ovid and PubMed platforoms. You can find these filters in the Cochrane Handbook (section 6.4.11)

Other pre-tested search filters for different types of experimental design and for different databases which you can copy and paste line by line to your search are also available from:

You can save search filters after you have typed them in by choosing to save your search history – then you can simply re-run them when you wish to apply them to a specific search. The filters look complicated but you should be able to cut and paste them into the search box in the database. Remember though – each line of the filter is a separate search on the database and you will need to combine them in exactly the same way as is set out in the filter before AND'ing them with your search strategy.

Publication Type limits in Databases

Some of the databases (including those on the Ovid platform, the HDAS NHS platform, and CINAHL) allow searches to be undertaken by publication or document type e.g. review, research, randomized controlled trial, editorial or letter. The actual publication types will differ across the different databases. Publication types, like subject headings, are assigned to each article at the point the article is added to the database.

Ovid Publication Types List - for Medline, Embase, PsychInfo etc.

CINAHL (EBSCO) Document Types List

NICE HDAS (for NHS) - for CINAHL, EMBASE, MEDLINE, PsycINFO etc. No full publication types list available (some examples are given in the 'About' Database link from the HDAS page). To locate the correct publication type you can either use one of the links above for the database you are interested in (as the terminology will be the same) or you can browse the thesaurus, typing in a publication type e.g. letter and then navigating to the broader terms within the thesaurus.

Some searchers use these publication types to limit their results to a particular type of trial design e.g. RCT, but remember that the pre-tested search filters mentioned on this page may be a more comprehensive way of limiting to trial design.

Other ways that publication types can be used is to limit a search (using a NOT command) to exclude e.g. letters, editorials etc. 

Example of a CINAHL NOT search

In most databases you have to type the NOT search line in as there is not a button to select to combine using NOT. The NOT command allows you to search for e.g. what is in the results of search line 1 which is not in the results of search line 2. It limits your results.

In the CINAHL example above you need to type in the relevant Search ID (see the column to the left of the Search Terms) around the NOT command.

On the OvidSP platform you simply type in the search line numbers around NOT e.g. 1 NOT 2.

 

Other limits - English Language, dates, full text.

Language limits. If time and money were no issue all systematic reviews would include all relevant research no matter what language it was originally published in. However, in many cases there are no funds for translation nor time to do this so it is fairly common that a filter is applied to limit results to just the English language (and this limitation is acknowledged in the methodology of the review). If you do apply an English language filter then do this at the end when you have finalised the search as you may decide, depending on the number of results, to examine relevant abstracts of foreign language papers as the abstracts are likely to be in English and may prove of interest even if you cannot include them completely.

Date limits should also always be used with care. In some instances they may be appropriate to apply e.g. if a drug has only been used to treat a specific condition from a certain date but prior to that was used for other conditions, but in many cases it is best to avoid date limits if at all possible as you are aiming to be as comprehensive as possible.

Full text limit option. This should be avoided. See the FUTON (Full Text On the Net) bias box on this page for more information.

Guidance on searching for systematic reviews - techniques

 

  • Use the MeSH Publication Type terms in Medline/PubMed:

    • META-ANALYSIS (since 1993)

    • REVIEW (since 1966)

       

  • Use the Emtree term in Embase:

    • SYSTEMATIC-REVIEW (since 2003)

 

  • Search filters to identify systematic reviews in Medline and other databases

     

    • The InterTASC Information Specialists’ Subgroup Search Filter resource website

 

Adverse effects - search filters

 

This following resource shows publications that have reviewed search filter performance and individual search filters:

Filters to Identify Studies of Adverse Effects

 

Other relevant articles:

The performance of adverse effects search filters in MEDLINE and EMBASE.

Golder S, Loke YK.

Health Info Libr J. 2012 Jun;29(2):141-51. doi: 10.1111/j.1471-1842.2012.00980.x. Epub 2012 Mar 14.

The contribution of different information sources for adverse effects data.

Golder S, Loke YK.

Int J Technol Assess Health Care. 2012 Apr;28(2):133-7. doi: 10.1017/S0266462312000128. Review.

Failure or success of electronic search strategies to identify adverse effects data. 

Golder S et al.

J Med Libr Assoc. (2012) 100:130-134

Poor reporting and inadequate searches were apparent in systematic reviews 

of adverse effects.

Golder S, Loke Y, McIntosh HM.

J Clin Epidemiol. 2008 May;61(5):440-8. doi: 10.1016/j.jclinepi.2007.06.005. Epub 2007 Oct 22.

 

Adding a search filter to your search: Example

Example of adding a filter to a search strategy - this is the 'Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity-maximizing version (2008 revision); Ovid format'.

1. randomized controlled trial.pt.

2. controlled clinical trial.pt.

3. randomized.ab.

4. placebo.ab.

5. clinical trials as topic.sh.

6. randomly.ab.

7. trial.ti.

8. 1 or 2 or 3 or 4 or 5 or 6 or 7

9. exp animals/ not humans.sh.

10. 8 not 9

 

 

  • The search strategy is copied and pasted in and then combined as directed by the filter.
  • To undertake a NOT search you simply need to type NOT in (this isn't one of the Combine options provided above the search box).
  • The final stage is to AND in the combined search filter to your search strategy - thereby limiting the results of the search to RCTs.

Saving Search Filters

If you copy and paste in the search filter(s) which you are going to use then it is a good idea to save them as a separate search strategy. You can then easily run this saved search e.g. for RCTs or for quantitative research and add it (AND it) to the search question you are currently researching without having to type the search filter in each time you wish to use it. 

FUTON (Full Text On the Net) bias

Some databases offer the option to limit to full text. Systematic Reviewers should not select this option as it will introduce bias into their research.

"Everyday information-seeking activities, especially by junior staff and students, often concentrate on research published in journals that are available as full text on the internet, and ignore relevant studies that are not available in full text, thus introducing an element of bias into their search result. [...] This bias may have the same effect in daily clinical practice as publication bias or language bias when doing systematic reviews of published studies". (Wentz, 2002).

In general if systematic reviewers locate articles which appear relevant at the abstract level but which they cannot access through full text subscriptions by their institutions they should try to source them through inter-library loans. King's staff (including NHS staff) and students all have an annual allocation of free inter-library loan requests. See the Interlending and Document Delivery pages for more information.

It is important to remember that if you select the limit to full text option in e.g. Ovid Medline you will only retrieve articles which are hosted on the Ovid platform/or which Ovid are aware are freely available as open access journals. You will exclude all other results even if King's has a subscription to them and they can be accessed by the SFX link. 

Wentz R. Visibility of research: FUTON bias. The Lancet. 2002;360(9341):1256. doi:10.1016/S0140-6736(02)11264-5