Searching for studies

Searching for studies

Review protocol

Early in the systematic review process a comprehensive search strategy will need to be developed. 

First, a MEDLINE  search is developed by the BMJT Information Specialist (Maria Clarke, email maria.clarke@manchester.ac.uk) in conjunction with the review authors.  The authors provide information on the population (Participant or Problem), interventions, comparisons and outcomes (PICO) for the development of the strategy.  Any relevant RCTs and/or a systematic review are also helpful for the testing of the strategy.  The agreed search strategy is inserted into the appendix of the protocol.

New Review

The development of the strategies for further database searches are based on the MEDLINE search strategy from the published protocol.

The following databases are searched for all Cochrane Reviews:

1.   Cochrane Central Register of Controlled Trials (CENTRAL)

2.   MEDLINE

3.   Embase

The following trial registers are also searched for each review:

1.       ClinicalTrials.gov

2.       WHO International Clinical Trials Registry Platform

 Searches of the above resources will take place at the editorial base and the results are provided as a text file, which can be imported into reference management software.

Searches must be no more than 12 months old when the review is published, therefore the searches of the main databases normally have to be re-run before publication.

 Structure of search strategies.  In databases such as, MEDLINE a search strategy would generally consist of three sets of terms a) the health condition, i.e. the population; b) the intervention; and c) the filter for the study design (usually a RCT filter).

RCT filters.  The RCT filter is used if there are a large number of results from a search (more than 500).  The RCT filter is added to the main search strategy with the aim of retrieving mostly randomised control or controlled trials.  You can find Information about filters in the Cochrane Handbook, section 4.4.7 and RCT filters developed for databases such as, MEDLINE can be found at Technical Supplement to Chapter 4: Searching for and selecting studies, section 3.6.

 Searching for and selecting studies.  Information to help with searching for and selecting studies can be found in Chapter 4 of the Cochrane Handbook.  The supplement Technical Supplement to Chapter 4: Searching for and selecting studies should be read in conjunction with Chapter 4. 

The minimum standards for searching for a Cochrane review can be found in the MECIR guidance

Additional searches are undertaken by the review authors, such as:

  1. Handsearching on subject-specific journals
  2. Searches of conference proceedings
  3. Checking of reference lists
  4. Contacting experts in the field.

Handsearching – “Handsearching involves a manual page-by-page examination of the entire contents of a journal issue or conference proceedings to identify all eligible reports of trials” (Cochrane Handbook, Section 4.S1: Technical Supplement to Chapter 4: 1.3.1 Handsearching). 

Sources to search – Information on sources to search is available in The Technical Supplement, section 1.3.  

Information on Identifying retracted publications (e.g. fraudulent publications), errata and comments can be found in Section 3.9.

Managing References.  Reference management software is helpful to keep track of references. For example: EndNote, Mendeley, RefWorks and Zotero

References can be imported into the reference management software using import filters, which will allow text files exported from various sources (e.g. CENTRAL, Medline) to be imported.