Finding Grey Literature for Systematic Reviews

Search grey literature for a systematic review by choosing three or four sources that fit the question, such as trial registries, theses, conference abstracts, and agency reports. For each source, record the date, the exact string or URL, the hits, how many you screened, how many were new includes, and the minutes spent. Put trial registries in the PRISMA 2020 registers box. Put websites, organisations, and citation searching in the flow diagram's other-methods column. Stop when extra sources add no new included studies. The worked log below is an illustrative example of how to measure that.

Author's note: My advice is always that academic databases are preferred and more reliable, specifically Medline (also known as PubMed), Embase, Web of Science, and Cochrane. These stand as the best and primary sources for systematic reviews. Grey literature searching is messy, complicated, and can quickly become overwhelming. A minor grey literature search is reasonable, but measure whether it is worth the effort and avoid being sunk into too much information.

Choose sources that are still available

Cochrane Handbook chapter 4 uses "grey literature" for reports published outside traditional commercial publishing. For Cochrane Reviews of interventions, searching reports, dissertations, theses, and conference abstracts is highly desirable (MECIR C28, section 4.3.5). The stated reason is to make the search extensive enough to reduce the risk of publication bias. Searching ClinicalTrials.gov and the WHO ICTRP search portal is a separate mandatory standard for those reviews (MECIR C27, section 4.3.3). The handbook says to run both, because ClinicalTrials.gov has search features the ICTRP portal does not. Build the database strategy first in the PubMed block translation guide.

Two sources from older lists are no longer what they were.

OpenGrey is a closed archive. GreyNet states that as of 1 December 2020 the OpenGrey repository was discontinued and preserved as a closed archive in the DANS EASY data archive. The same notice says GreyNet's own OpenGrey content was transferred to the GreyGuide repository. That is GreyNet's material, kept after the closure. GreyNet does not name a live database that replaces the old OpenGrey search.

EThOS, the British Library service for UK doctoral theses, was restored on 8 July 2026. The British Library notice says the restored platform provides access to over 650,000 metadata records and does not hold full-text thesis files. Almost 65% of those records link to an institutional repository, where the full text can be consulted if the university has made it available. If a record has no repository link, the Library says to contact the awarding university.

Agency reports still need a named site. WHO IRIS is one live example. Pick the agency that actually publishes on your question, and log that URL.

Worked search log

The question, strings, counts, and minutes below are invented. They show the columns. They are not the results of a real review.

Illustrative question: do community exercise classes reduce falls in adults aged 65 and older?

Source	Date searched	String or URL	Hits	Screened	New includes	Minutes
ClinicalTrials.gov	4 Mar 2026	falls AND exercise	12	12	1	25
WHO ICTRP	4 Mar 2026	falls AND exercise	9	9	0	20
Agency reports page	5 Mar 2026	site:who.int falls AND exercise filetype:pdf	6	6	1	40
Reference lists of the two includes	5 Mar 2026	backward citation check of the register record and the agency report	4	4	0	15
Email to one falls charity	6 Mar 2026	request for unpublished evaluations	1	1	0	20

Minutes in the example: 25 + 20 + 40 + 15 + 20 = 120. New includes: 2. That is 60 minutes per new include. The ratio is arithmetic on these made-up rows, so you can see the cost of the extra includes in one review. It is not a published benchmark.

The example stops after the reference-list check and the charity email. Those were the last two sources, and both added zero new includes. Cochrane and PRISMA ask you to report the sources you used. This two-source end point is only the rule inside the example. If time is the hard limit, rapid review methods treats a limited grey search as a stated cut in the protocol.

Published time data are a description of one sample, not a budget to copy. Saleh, Ratajeski, and Bertolet (Evidence Based Library and Information Practice, 2014) reported a median of 85 minutes searching grey literature and a median of 471 minutes searching all resources. In the discussion, grey-literature searching was 27% of total search time on average, and half of the participants spent 20% or less. The results section says 17 of 81 respondents (21%) completed both parts of the study and were included in the analysis. Use your own log for the review in front of you.

Where each row goes on the PRISMA 2020 flow diagram

Rethlefsen and Page (Journal of the Medical Library Association, 2022) describe the PRISMA 2020 flow diagram for new reviews that searched databases, registers, and other sources. Column 1 is identification via databases and registers. Column 2 is identification via other methods: websites, organizations, and citation searching. Template files are on the PRISMA 2020 flow diagram page.

PRISMA-S (Rethlefsen et al., Systematic Reviews, 2021) is the search-reporting extension. Item 3 asks you to list study registries. Item 4 asks you to describe online sources you purposefully searched or browsed, and how. Item 5 asks you to describe citation searching. Item 6 asks you to say whether you sought studies or data by contacting authors, experts, manufacturers, or others.

Example row	Flow diagram field	PRISMA-S item
ClinicalTrials.gov and WHO ICTRP	Registers, in the databases and registers column	3. Study registries
Agency reports page	Websites, in the other-methods column	4. Online resources and browsing
Reference lists	Citation searching, in the other-methods column	5. Citation searching
Charity email	Organisations, in the other-methods column	6. Contacts

In the PRISMA flowchart generator those other-methods fields are labelled Websites, Organisations, and Citation chasing. A databases-only path, from export totals to records screened, is the worked example in from deduplication counts to a PRISMA flowchart. Add register and other-methods counts when the log has them. Deduplicate grey records together with the database exports before the screened total goes into the figure. The match order and the audit are in deduplicating multi-database search exports. Records saved as RIS can enter screening through the Zotero, EndNote, and RIS import.

A log with a date and a string for every source is what makes the search report aligned with PRISMA 2020. The flow diagram and the log have to name the same sources.

Appraise the documents you keep

Tyndall's AACODS checklist (Flinders University, November 2010) is a grey-literature appraisal checklist: Authority, Accuracy, Coverage, Objectivity, Date, and Significance. The same checklist says unpublished studies and randomised trials should be appraised with the same tools as their published counterparts. Use AACODS for a report or other grey document. Use the review's usual risk-of-bias tool when the record is a trial.