The PRISMA 2020 checklist (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) is a 27-item evidence-based reporting standard designed to ensure transparency across literature searches, study selection, risk-of-bias evaluation, and statistical synthesis.
What Changed in the PRISMA 2020 Statement?
The updated PRISMA 2020 framework reflects contemporary systematic review practices, including automated machine learning screeners, pre-registration registries (such as PROSPERO), open science data sharing, and updated risk-of-bias tools (like RoB 2 and ROBINS-I).
PRISMA is not a quality appraisal instrument for the underlying studies; it is an author transparency tool ensuring reviewers provide enough procedural detail for independent replication.
The 27 Checklist Items Across Core Sections
To satisfy peer review requirements in high-impact medical and scientific journals, structure your manuscript across PRISMA’s 7 primary reporting domains:
| Section | Checklist Items | Essential Reporting Requirements |
|---|---|---|
| Title | Item 1 | Identify the report explicitly as a systematic review, meta-analysis, or both. |
| Abstract | Item 2 | Follow the structured PRISMA 2020 for Abstracts (background, objectives, eligibility, sources, risk of bias, synthesis, registration). |
| Introduction | Items 3–4 | Describe the scientific rationale and explicitly define review objectives using the PICO framework (Population, Intervention, Comparator, Outcome). |
| Methods | Items 5–15 | Report databases, full reproducible search strings, screening criteria, data extraction tools, risk-of-bias assessment methods, and statistical synthesis models. Read How to Write the Methods Section. |
| Results | Items 16–22 | Document search results, study characteristics, individual study risk of bias, synthesized outcomes, and sensitivity analyses. Read How to Write the Results Section. |
| Discussion | Item 23 | Provide evidence interpretation, discuss limitations of the included evidence and review process, and state real-world implications. Read How to Write the Discussion Section. |
| Other Info | Items 24–27 | State protocol registration details (e.g., PROSPERO ID), funding sources, competing interests, and open data/code availability. |
The 4 Stages of the PRISMA 2020 Flow Diagram
A central requirement for systematic reviews is the standardized PRISMA flow diagram tracking the lifecycle of identified records:
- Identification: Total records identified across databases (PubMed, Scopus, Web of Science, Embase), grey literature, and citation registers, alongside duplicates removed.
- Screening: Records screened at title/abstract level and the exact count of records excluded based on pre-defined criteria.
- Eligibility: Full-text articles retrieved and evaluated, detailing the explicit reason for exclusion for each individual paper.
- Included: Final count of studies included in qualitative synthesis and quantitative meta-analysis.
Common PRISMA Reporting Errors That Cause Rejection
- Omitting Reproducible Search Strategies: Reporting keywords without providing the exact boolean search strings used in at least one major database (typically placed in the Supplementary Materials).
- Flow Diagram Arithmetic Mismatches: The number of records in the flow diagram not reconciling with figures in the Results section.
- Missing Protocol Registration: Failing to report whether a formal protocol was registered on PROSPERO, OSF, or Cochrane prior to screening.
- Unexplained Discrepancies: Not specifying whether screening and data extraction were performed by two independent reviewers to prevent bias.
Systematic Review Author Audit Checklist
- The manuscript title contains “Systematic Review” and/or “Meta-Analysis”.
- Protocol registration number (e.g., PROSPERO) is cited in the Abstract and Methods.
- At least one comprehensive, line-by-line database search strategy is included in full.
- Independent double-screening and data extraction procedures are documented.
- The validated PRISMA 2020 Flow Diagram is generated and arithmetic matches the text.
- Validated risk-of-bias instruments (RoB 2, ROBINS-I, Newcastle-Ottawa) are reported.
- Data sharing, code availability, and funding declarations are transparently stated.