The PRISMA flow diagram and checklist
The flow diagram is computed from the records, so its counts always reconcile. The drafted methods text and the 27-item checklist come from the same data.
PRISMA 2020 is the reporting guideline for systematic reviews. IAMB.AI produces its two main pieces from your records.
The flow diagram
The PRISMA tab draws the diagram from what the review recorded:
- identified: records from each database search;
- removed before screening: duplicates (and records marked ineligible by automation tools, if the AI auto-excluded);
- screened: records judged at title and abstract, and how many were excluded;
- sought for retrieval and not retrieved;
- assessed for eligibility at full text, with the excluded reports per reason;
- included in the review.
The counts always reconcile
| Check | Rule |
|---|---|
| Identification | identified = screened + removed |
| Screening | screened = excluded + sought + pending |
| Eligibility | assessed = excluded + included + pending |
Because every number comes from the records, the diagram, the methods text and the paper agree. Records still waiting for a decision show as pending until the work is done.
Drafted methods text
Next to the diagram, the tab drafts the methods paragraphs that describe the searches, deduplication, screening (including whether it was blinded and the agreement), retrieval and eligibility, with the numbers filled in.
The checklist
The Checklist tab is the 27-item PRISMA 2020 checklist, filled in with where each item is reported in your paper. Journals usually ask for it with the submission.
PRISMA items the app helps with directly include item 8 (selection process), item 11 (risk of bias assessment, through the quality checklist), item 18 (risk of bias in studies, the quality table) and item 24 (registration).