Systematic and systematic-like review toolkit

Filters and complex search examples  

Database built-in search limits

Many databases include built-in search limiters that can be applied to limit search results. These vary by database and common examples include limiting by age group, publication type, publication dates, study type, publication language, and more.
 

Note

Before using a database’s built-in limit in your final search, it is recommended to experiment with how it impacts on the results. Using database limits may save you time, but care should be taken to avoid inadvertently excluding relevant articles.

Potential issues include:

•  recent studies that have not been indexed against the limiter will not be retrieved
• relevant articles may not use the keywords that the database expects.

Generally, the use of these limiters in advanced reviews is discouraged.  

Justifiable limits

Use of any database limits in your search, including language and date, needs to be justifiable.  All limits should be detailed in the protocol and well established before performing the search.  

Limits such as publication date and language limits should be applied consistently to all searches and translations. Due to variations in database topic and study design limits, there may be justifiable reasons why a limiter is used in one database but not in another.
 

Tip

A commonly used database limit is "English Only". or adding the language to your search terms (e.g., "English"). However, it is preferable to remove these articles later during the screening process. This will ensure you do not miss articles improperly indexed or published in multiple languages (e.g., English and German").

Search filters

A search filter (also known as a search hedge) is a pre-made search strategy. Typically developed to search complex concepts (or concepts that require many search terms), search filters can be copied and pasted into database search boxes and combined with the rest of your search.  

The use of a search filter can save time and effort, in both the creation of the overall search strategy and in significantly reducing the number of records that researchers need to screen. Search filters can be designed and tested to be either:

• Precise - to return as few irrelevant records as possible, but this also has high potential for missing relevant literature.
• Balanced between sensitivity and precision - aiming to retrieve as few irrelevant records as possible.
• Specific - optimal for use when requiring a strategy that may not identify all papers, however, will retrieve the most relevant papers.

Researchers should investigate to see if the filter is high or low sensitivity and precision to determine whether to use it or not.

There are a range of search filters available. Some are designed to retrieve specific study designs such as randomised controlled trials or systematic reviews.  Other types of search filters will retrieve records based on topic, or to identify clinical concepts, such as diagnosis, etiology, prognosis and treatment, or features such as the age or gender of research study participants.
 

Caution

While some filters have been validated for their effectiveness in retrieving relevant results, many have not. These may have been tested but not validated for sensitivity, specificity, precision, and accuracy. Validated filters have been tested against gold-standard sets of bibliographic records.
 

Factors to consider when using a search filter include:

• How current is the filter? Does it reflect language and subject heading changes? Field codes also change over time.
• Does the filter consider local terminology, for example if the filter was created within US context but is needed for an Australian search.  
• Who was it created by? Was an expert information manager/librarian involved?
• What methods were used to develop it?
• Is there information available on if/how the search filter was tested? It is always worthwhile to proofread any filters you consider using in your search as errors or typographical errors may be present.
   

Note

Most search filters are only available for certain databases, and are interface-specific, due to the different syntax and controlled vocabulary used across databases. Search filters are not available for all databases or all database interfaces.

It is recommended to experiment with how a filter can impact on the search results before finalising your search.

---

Editing filters

There may be occasions where a search filter isn’t quite what you are after. You can edit a search filter to fit your search needs.  

When translating your search from one database to another, there may not be an equivalent search filter available. For example, there may be a topic filter for use in PubMed, but not for Embase. The use of an (edited) search filter in a database it was not designed for may mean that the results are not the same as was intended by the filter's creator(s).

Translation or adaptation of search filters from one interface to another can be complex and time consuming and should be undertaken carefully. 
 

Tip

If you use a search filter in your search, this needs to be cited. Also, if you use, but edit a filter, this needs to be noted in your paper with an indication of why you made the change(s) and include an acknowledgement of the original researchers.

---

Critical appraisal of filters

The development of search filters has become more objective and rigorous over time.  Published search filters should detail the methods used by the authors, which can provide guidance for determining the suitability of a search filter.

In addition, the quality of search filters can be appraised using critical appraisal tools.  These tools can assess the focus of the filter, the methods used in its creation, as well as the quality of the testing and validation which have been conducted.
 

Note

The use of search filters does not include an assessment of the quality and relevance of evidence in records retrieved in the search.

Sources of filters

Note

Inclusion of a search filter is not an endorsement of its validity or a recommendation. 

CADTH 

Filters are designed for PubMed, OVID Medline, Embase, PsycINFO, for systematic reviews/meta-analysis/health technology assessment, randomised controlled trials, guidelines, economic evaluations/cost/economic models, health utilities/quality of life, and narrow economic filter. 

Center for Evidence-Based Management (CEBMa) 

Filters to identify meta-analyses and systematic reviews and controlled and/or longitudinal studies.  

Cochrane Technical Supplement: Search Filters 

Includes filters to identify randomised controlled trials in several databases. 

Flinders Filters (including CareSearch Filters) 

Filters available on subjects including Aboriginal and Torres Strait Islander Health, Bereavement, Contraception, Dementia, Heart Failure, Grief and loss, Integrated care, Lung Cancer, Ophthalmology, Palliative Care; Primary Health Care; Residential Aged Care; and Stroke. 

ISSG search filters resource 

Large resource of filters on topics including diverse effects, etiology, diagnostic studies, economic evaluations, epidemiological studies, guidelines, mixed methods studies, non-randomised studies, observational studies, outcome studies, prognosis, public views/patient issues, qualitative research, quality of life, quasi-experimental studies, RCTs/other trials, systematic reviews, therapy studies, adherence, age, animal studies, clinical examination, EBP, gender, geography, health care delivery, health equity, patient safety, population groups, and quality improvement.  

There is also guidance on how to critically appraise search filters, as well as information on the development and use of search filters. 

McMaster 

Filters designed for use in MEDLINE, EMBASE, and PsycINFO in categories of therapy, diagnosis, review, prognosis, causation (etiology), economics, cost, clinical prediction guides, and qualitative studies. 

NSQHS Standards: live literature searches, tested and endorsed as part of the Health Libraries for the National Standards Research Project (HeLiNS), a joint initiative of Health Libraries Australia and Health Libraries Inc. These automatic PubMed searches are focused on 12 topics of relevance to best practice in Australian hospital settings and are available in subsets including the last 5 years, limited to Australia and limited to Reviews.    

Ovid Expert Searches 

Filters for used in Medline and Embase on a range of health topics, diseases and treatments, including some COVID, Zika, Ebola, Gender, Age, Pharmacovigilance, Family Medicine, and Mixed Methods. Note most are not validated. 

SIGN (Scottish Intercollegiate Guidelines Network) search filters 

Filters are designed for use in Medline, Embase and CINAHL on the Ovid platform, along with other more specialised databases. Filters include systematic reviews, randomised controlled trials, observational studies, diagnostic studies. There are also economic studies and patient issues filters.  

Topic search filters: a systematic scoping review 

This scoping review identifies 58 filters across a range of topics. 

University of Alberta Library – Health Sciences Search Filters 

Geographic Filters, Indigenous Peoples - including Australasia, Socio-Economic Factors/Determinants (e.g., Poverty, Religion, Food Security), Specific Conditions (e.g., Obesity/Caregiver Burnout), Subject Filters (Environment/Pollution, Climate Change), Methodological Filters, Population Groups (e.g., Pediatrics), Settings (e.g., Emergency Department, Telehealth), Therapy/Interventions (e.g., Exercise, Cannabis).