What PRISMA is and who maintains it
PRISMA, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, is a 27-item checklist and an accompanying flow diagram for reporting systematic reviews, including reviews that statistically combine results as a meta-analysis. The current version, PRISMA 2020, was developed by an international group of methodologists and updated the original 2009 statement to reflect advances in review methods (Page et al., 2021).
The checklist covers the title and abstract, the rationale and objectives, eligibility criteria, information sources and search strategy, the process for selecting and extracting data, the methods used to assess risk of bias and certainty of evidence, and how results were synthesised and presented. The flow diagram tracks records from identification through screening, eligibility assessment and inclusion, showing exactly how many records were excluded at each stage and why.
Why it exists
A systematic review is meant to be a reproducible method for answering a research question from the existing literature, not a list of studies the author happened to find and like. Readers, including other researchers who will later update the review, need to know exactly what was searched, how studies were screened, which were excluded and why, and how the included studies were combined. Without this, two reviews on the same question can reach different conclusions for reasons a reader cannot detect.
PRISMA exists to make this process visible. The EQUATOR Network, which hosts PRISMA alongside other reporting guidelines, was itself set up because incomplete reporting across many types of health research was recognised as a problem that no single guideline could solve on its own, and a coordinated library of guidelines was needed (Moher et al., 2008).
There is also a narrower, more immediate reason PRISMA matters to the person writing the review. A reviewer who records search dates, screening decisions and exclusion reasons as they go, in the format PRISMA expects, rarely has to reconstruct that history under pressure near a submission deadline. The discipline the checklist imposes during the review is at least as valuable as the document it produces at the end.
Which reviews it fits
PRISMA applies to systematic reviews, with or without meta-analysis, across any field, not only medicine. A review counts as systematic when it has a protocol defined before searching (ideally registered, for example with PROSPERO), explicit eligibility criteria, a documented and reproducible search of one or more sources, and a defined method for selecting, appraising and synthesising the studies found. Registering the protocol before searching serves the same purpose here as preregistration does for primary studies: it fixes the review question and inclusion criteria before the results of the search are known (Nosek et al., 2018).
| Review type | Use PRISMA? | Notes |
|---|---|---|
| Systematic review (no meta-analysis) | Yes | Narrative or tabular synthesis still needs the full checklist |
| Systematic review with meta-analysis | Yes | Additional items on statistical methods, heterogeneity and certainty apply |
| Scoping review | Use the PRISMA extension for scoping reviews (PRISMA-ScR) | Different objective: mapping a field, not answering a focused question |
| Review protocol (registered before the review) | Use PRISMA-P | A separate, earlier-stage checklist for the protocol itself |
| Narrative or traditional literature review | No | State plainly that it is not systematic; do not claim PRISMA compliance |
| Umbrella review of other systematic reviews | Use the PRISMA extension for overviews | Distinct items for assessing and comparing included reviews |
Where PRISMA does not apply
Not every doctoral literature review is, or needs to be, a systematic review. A review that sets out to situate a thesis within a field, trace the development of a theory, or build a conceptual framework is a legitimate and common doctoral task, but it is not answering one focused question through a reproducible search and appraisal process. Claiming PRISMA compliance for this kind of review overstates what was done; it is better to describe the approach honestly as a narrative or critical review and explain how sources were chosen.
PRISMA also does not assess the quality of evidence within included studies on its own. Risk of bias in individual studies and the overall certainty of a body of evidence are separate steps, commonly carried out with tools such as GRADE, that PRISMA reporting should disclose rather than replace (Guyatt et al., 2008).
Finally, PRISMA items that assume a single, answerable question sit awkwardly on a review with a genuinely exploratory or mapping purpose. Trying to force a scoping exercise into PRISMA 2020's items on effect synthesis and heterogeneity, rather than using the scoping review extension, usually produces a report that looks complete but answers a question the review was never designed to answer.
Why reporting quality matters here specifically
A systematic review is often used as a single, authoritative summary of a whole body of research, so errors or omissions in how it was conducted propagate further than they would in a single primary study: a clinician, policymaker or later meta-analyst may rely on the review's conclusion without re-examining every included study themselves. Incomplete reporting in observational primary studies has been one of the stated reasons general reporting guidance is needed at all (von Elm et al., 2007), and the same logic applies, at a larger scale, to the review that summarises those studies: what the review does not disclose about its own search and selection process cannot be checked by anyone reading it.
How it is applied in practice
Most health and social science journals that publish systematic reviews require a completed PRISMA checklist, with page references, submitted alongside the manuscript, and many ask authors to register the review protocol on PROSPERO or a similar registry before screening begins. Reference management and screening software commonly includes built-in PRISMA flow diagram generators that count records at each stage automatically, which removes a frequent source of error: reviewers manually recalculating numbers after the fact and losing track of why a given record was excluded.
Software support has made the mechanics of PRISMA reporting much less burdensome than it was when the original 2009 statement was published; even so, the software can only count what the reviewer logs, so the underlying discipline of recording decisions as they happen still rests with the researcher.
For a doctoral researcher, the practical sequence is to register the protocol, keep a running record of every database searched with its date and search string, log exclusions with reasons as screening happens rather than reconstructing them afterwards, and only then draft the PRISMA checklist and diagram from that record. Depositing the extracted data and, where used, the analysis code alongside the published review follows the same FAIR principles of findable, accessible, interoperable and reusable outputs that apply to primary research data (Wilkinson et al., 2016).
An illustrative example
The passage below is a hypothetical illustration, not a quotation from a published review.
| Element | Without PRISMA | With PRISMA |
|---|---|---|
| Search | "A literature search was conducted in several databases." | "MEDLINE, Embase and CINAHL were searched from inception to 14 March 2026 using the terms in Appendix A; reference lists of included studies were also checked." |
| Selection | "Studies were screened for relevance." | "1,204 records were identified; after removing 318 duplicates, 886 were screened by title and abstract, 94 assessed in full text, and 31 included, with reasons for exclusion recorded for each stage in a PRISMA flow diagram." |
| Risk of bias | Not mentioned | "Risk of bias in each included study was assessed independently by two reviewers using the Cochrane RoB 2 tool, with disagreements resolved by a third reviewer." |
| Synthesis | "Results were broadly consistent across studies." | "A random-effects meta-analysis was performed; heterogeneity was assessed with I-squared (62%, indicating substantial heterogeneity) and explored in a pre-specified subgroup analysis by study setting." |
| Certainty | Not addressed | "The overall certainty of evidence for the primary outcome was rated moderate using GRADE, downgraded for inconsistency." |
What to check before you submit
- Is the review protocol registered, and does the manuscript say so with a registration number?
- Does the flow diagram show counts at every stage, including duplicates removed and exclusions with reasons?
- Are the databases, dates and full search strings reported, ideally in an appendix?
- Was risk of bias assessed in each included study with a named tool, by more than one reviewer?
- If results were combined statistically, is the method, the heterogeneity statistic and any subgroup or sensitivity analysis reported?
- Is the certainty of evidence for each main finding stated, and is the basis for that rating given (Guyatt et al., 2008)?
- Official checklist and source: PRISMA Statement (prisma-statement.org) .
Key takeaways
- PRISMA 2020 is a 27-item checklist and flow diagram for reporting systematic reviews and meta-analyses (Page et al., 2021).
- It exists because a review's conclusions can only be checked if its search, screening and synthesis are reported completely.
- It fits systematic reviews, including meta-analyses; a narrative or conceptual literature review should say plainly that it is not systematic.
- PRISMA reports a review's process; GRADE separately rates the certainty of what it found (Guyatt et al., 2008).
- EQUATOR hosts PRISMA alongside the wider family of reporting guidelines it was set up to coordinate (Moher et al., 2008).
