Full-Text Screening and Exclusion Reasons

From abstracts to full texts

Full-text screening is the second selection stage: every record you marked Include or Maybe at the title and abstract stage is re-assessed against the eligibility criteria using the complete paper. In the tool, click Full-text screening in the sidebar (a Pro feature on a cloud-saved review). The panel lists exactly the records you kept; if it is empty, screen titles and abstracts first.

Get the full text

For each record you can upload the PDF (click Upload PDF) and read it in the built-in viewer without leaving the tool. Before hunting through your library, check the open-access finder: when a free, legal copy exists, the panel shows an Open the free PDF button sourced from Unpaywall. If no free full text is found, the tool says so and suggests your library or a manual upload. Lookups are cached on the review, so a checked record is never queried twice.

Batch helpers save the clicking on large reviews. Find open access for all runs the Unpaywall lookup for every record with a digital object identifier that has no PDF yet and reports how many free copies it found. Bulk upload PDFs takes a whole folder of PDFs at once and matches each file to its record by digital object identifier or title in the filename, then shows a report of what was attached, what failed, and which files matched no record.

Find open access PDFs lists the free copies and can attach them to your records in one step. A copy that cannot be attached stays as a link you can open.

You can also drop a PDF straight onto the panel to attach it to the selected record, or drop several at once to send them to the bulk matcher.

For anything still missing, Library request gathers those records into an interlibrary loan CSV, a RIS file, and email text for your librarian. Run Find open access for all first so you do not request papers that are free.

Dual (blinded two-reviewer) full-text screening

The owner or an arbiter can switch on dual screening in the stage settings. Each reviewer then votes include or exclude (with a reason) without seeing the other reviewer's vote. Matching votes decide the record automatically. Disagreements, including two excludes with different reasons, appear in the Conflicts view, where the owner or an arbiter records the final consensus decision; PRISMA counts always follow the consensus decision, never an individual vote.

The stage settings also let you replace the built-in exclusion-reason list with your review's own reasons, so the reasons in your PRISMA flow diagram match your protocol word for word.

Decide and record the reason

1. Select a record in the full-text list.
2. Read the PDF in the viewer, or open the free full text.
3. Click Include if the study meets all criteria.
4. Click Exclude if it fails any criterion, then pick the exclusion reason. The built-in reasons are: Wrong study design, Wrong population, Wrong intervention, Wrong comparator, Wrong outcomes, Wrong setting, Wrong publication type (conference abstract, protocol, etc.), Full text not available, Not in eligible language, Duplicate / companion report, and Retracted. You can also type a custom reason. If you press Exclude with no reason chosen, the panel asks why and one tap on a reason completes the exclusion.
5. Move to the next record. Decisions save as you go and are logged in the audit trail.

PRISMA counts flow automatically

PRISMA 2020 requires you to report how many reports were sought, assessed at full text, and excluded with reasons. The tool tracks all of it: full-text decisions and their reasons feed the flow diagram builder, where reports assessed, exclusions grouped by reason, and final included counts appear in the right boxes. Open Build PRISMA diagram in the sidebar to see the figure update, adjust any number, and export it as an image for your manuscript.