Editorial and Publishing Policy Resource
Reporting search dates in Cochrane Reviews

Update, January 2015: At the joint meetings of the Co-ordinating Editors, Managing Editors, and Trials Search Co-ordinators held during the Cochrane Colloquia in Auckland (2012) and Quebec (2013) it was agreed that ‘Date of search’ should be used instead of ‘Assessed as up-to-date’ to reflect the search and full incorporation of all search results into a Cochrane Review. The technical issues associated with retiring the ‘Assessed as up-to-date’ field are currently being addressed by Cochrane and Wiley; however this will not prevent editorial teams from adhering to this new policy on how to report search dates.

One date should be used to reflect the search and full incorporation of all search results into the review; this date is the ‘Date of search’ and should be highly visible to users/readers. Standard practice has been to publish ‘Assessed as up-to-date’ field and not the ‘Date of search’. Until the ‘Assessed as up-to-date’ field is removed these two dates must be the same.  

Definitions of search types (full, top-up, scoping) are available in Table 1.

  1. MECIR conduct standard C37 requires that searches for all relevant databases be run (or re-run) within 12 months before publication of the review or review update, and that the results are screened for potentially eligible studies.
  2. Updates vs. amendments: a review is considered updated and receives a new citation in Cochrane Database of Systematic Reviews (CDSR) when a new search is conducted and the results of the search are fully incorporated. If a scoping search is conducted to determine if an update is required, then the date of this search will not change the ‘Date of search’ in the review or lead to a new citation version being created. This should be published as an amendment if necessary (see guidance).
  3. If top-up searches are performed and the results incorporated then that top-up search date becomes the date of the full search (i.e. the date that appears in the ‘Date of Search’ field).
  4. If top-up searches are performed and the results are NOT fully incorporated then: 
    1. The ‘Date of search’ remains the date of the search for which results were fully incorporated.
    2. Studies not yet fully incorporated into the review are added to ‘Studies awaiting classification’.
    3. The 'Search methods' in the abstract should focus on reporting the search dates related to the last fully incorporated studies. Brief mention of a top-up search may be made only if it was conducted for a completed update or new review. Do not refer to scoping searches for updating in the abstract.  
    4. The ‘Search methods for identification of studies’ in the main text of the review should be used primarily to describe the details of the search for which the results have been fully incorporated, i.e. the dates of individual database searching and the hits retrieved should be based on the search date where results are fully incorporated. If a top-up search has been performed, but the results not yet fully incorporated, the search section may briefly describe this (see example below) and state how many studies have been placed in ‘Studies awaiting classification’. Further details of the top-up search, such as the dates each database was searched may be given in an appendix. 
    5. In the 'Results of the search' section the authors should specify the number of studies yet to be fully incorporated into the review. This should also be reflected in the conclusions (both of the main review and the abstract).
    6. The PRISMA flow of studies diagram should also reflect the number of studies in the ‘Studies awaiting classification’ section.
    7. The 'What’s New' events must describe the number of studies that have been put into ‘Studies awaiting classification’ if the top-up search is mentioned in the search methods section.
  5. If different databases were searched on different dates, the most recent date of the search for each database should be given within the text of the review and the earliest of these dates should be entered as the ‘Date of Search’. In the case of review updates or ‘top-up’ searches, if there is clear rationale for not searching one or more of the previously searched databases (e.g. because no unique relevant records were identified in the original/previous search, or the database is no longer being updated), the rationale should be stated within the text of the review. In this case, the ‘Date of Search’ should be the earliest date of the searches performed for this smaller set of databases.

Table 1. Definitions of search types (full, top-up, scoping)

Search types

Definition

Full search – results fully incorporated

Electronic search strategies run in full in all relevant databases AND all search results are assessed for eligibility and either included or excluded

Top-up search – results not fully incorporated 

Electronic search strategies run in full in all relevant databases BUT search results are not all assessed for eligibility, instead they are placed in Studies awaiting classification

Scoping search for updating

Electronic search strategies run in selected databases to determine if an update is required


Examples of reporting top-up searches

The number of instances where a top-up search is performed and potential new studies are identified but not fully incorporated before publication should remain low. The following examples show how such searches should be described in various sections of a systematic review:

What’s New

Do not change the ‘Date of search’ or the ‘Assessed as up-to-date’ in the ‘Review information’ section. Also, if less than 10 trial reports then list here in parentheses and link. For example:

“The search was updated in month/year and n trial reports added to ‘Studies awaiting classification’ (e.g. Bertini 2005; Crowther 2005; Gillen 2004).”

Abstract

Search methods 

The focus should remain on the text about previous searches (fully incorporated) but the top-up search may be mentioned. For example:

“We searched the Cochrane Central Register of Controlled Trials, MEDLINE, Embase, and CINAHL (June 2013). We updated this search in September 2014, but these results have not yet been incorporated.” 

Main text

Search methods for identification of studies 

The search should be reported as per MECIR reporting standards R34 to 39, including the dates for each source. At the end of the search methods section, it is appropriate to add the following text:

“We performed a further search in [month/year]. Those results have been added to ‘Studies awaiting classification’ and will be incorporated into the review at the next update.”

Do not list all databases and the dates. If a top-up search in reported in this section, only a single month (or range of months) and year should be shown.

Results: Description of studies

This section will differ depending on the review, so add text where it is most appropriate); for example: 

“[insert number] study reports from an updated search in [month/year] have been added to ‘Studies awaiting classification’.”

Discussion: Potential biases in the review process

Acknowledge the potential impact of un-incorporated studies as a source of potential bias, especially if studies concerned are potentially important in terms of sample size or direction of effect; for example:

“We attempted to conduct a comprehensive search for studies, but the fact that [insert number] studies have not yet been incorporated may be a source of potential bias.”

Authors’ conclusions (Implications for practice)

This is not an implication for practice as such, but users should be alerted to the issue of un-incorporated studies, particularly if the studies concerned are potentially important in terms of sample size or direction of effect; for example:

“The [insert number] studies in ‘Studies awaiting classification’ may alter the conclusions of the review once assessed.”

Section info
Describe change
Added text to describe reporting search dates "if different databases were searched on different dates".

Section 4.1 - removed "(and this date must be within a year of publication, and ideally within six months).” - [requested by RF]

Change date
1 September 2016