Grey Literature Search Checklist and Log for Systematic Reviews

Make partly reproducible searching honestly auditable

Grey literature can include trial registers, repositories, theses, regulatory material, government and organizational reports, conference outputs, and other evidence outside conventional journal publishing. The appropriate sources depend on the review question; there is no universal list.

Grey-literature searching may identify evidence affected by publication or dissemination processes, but source instability, weak search interfaces, personalization, changing indexes, and incomplete metadata can limit reproducibility. A defensible log records not just the query, but also the source, route, date, screen depth, stopping rule, yield, retained records, storage, contacts, and limitations.^1,2

What the log must preserve

1. Why this source: coverage, audience, evidence type, geography, or regulatory role.
2. How it was searched: exact query, filters, browse path, or contact method.
3. How far: result pages, records, folders, years, or another screen-depth rule.
4. When and by whom: date, time when relevant, searcher, account or location context if material.
5. What happened: yield, records exported or captured, unique/retained records, duplicates, and file location.
6. What cannot be reproduced: ranking, personalization, dynamic content, unavailable history, or undocumented changes.

Dynamic sources require more than a query box

Bibliographic databases are not the only routes to eligible evidence. Trial registers may contain unpublished or ongoing studies; repositories may preserve theses and reports; regulators and health-technology-assessment bodies may publish evaluations; organizations may host technical documents; and contact with experts may identify material that is poorly indexed. The value of each source depends on the review’s eligibility criteria and the publication pathways of the evidence—not on whether it appears on a generic checklist.^3,4

Many grey-literature interfaces are difficult to reproduce exactly. Rankings can change, websites are revised, results may depend on location or account state, search histories may not be exportable, and browsing may be more effective than a formal query. Transparent documentation does not pretend these limitations disappear. It records enough context for another researcher to understand what was attempted, how far the search proceeded, what was retained, and where uncertainty remains.^3–5

Define the stopping rule before fatigue defines it

A stopping rule makes screen depth explicit. It might specify a number of result pages, a ranked-record limit, a date boundary, a set of folders, saturation after consecutive pages without a potentially relevant record, or completion of a named organizational pathway. The rule should fit the source and be documented prospectively where possible. It is a pragmatic boundary, not evidence that nothing relevant exists beyond it.

Preserve the route, not only the destination

When browsing, record the sequence of menus, collections, filters, alphabetic lists, or site sections. When using a search engine, record the exact query, date, result depth, and any material context such as language or location. When contacting an organization or expert, preserve the recipient role, request, date, response, attachments, and follow-up. A citation found at the end is not enough to reconstruct the discovery method.

Separate discovery from appraisal

Grey-literature status does not determine study quality. The search log records discovery and handling; eligibility assessment, risk-of-bias assessment, and certainty judgments occur in their own review processes. Similarly, finding non-journal material may reduce some publication or dissemination bias in a particular evidence base, but it cannot remove all bias or demonstrate that the search is comprehensive.^6,7

Record uncertainty as data

Dynamic searching becomes more trustworthy when limitations are recorded alongside the result. A reproducibility flag should explain whether the source provided a stable query history, whether rankings or content could change, whether all screened records were exportable, and whether the search could be rerun by another person. This is not a pass/fail score. It is information needed to interpret the method and plan updates.

The log below converts this ladder into fields that can support PRISMA-S reporting and a future update. Complete only relevant modules, but never leave the source rationale, executed route, screen boundary, disposition, and limitation status implicit.

Use the log with its methodological guide

The log captures what happened. The paired guide explains when grey-literature searching is warranted, how to select sources, and how to report limitations without overstating coverage.