Prevention & care
Systematic reviews and meta-analysis: review purposes, pooling, overlap and interpretation
On this page
A systematic review uses explicit methods to assemble and synthesize studies addressing a defined question. Meta-analysis is a statistical way of combining results, rather than a synonym for the review itself. PRISMA separates these terms and asks authors to report their methods transparently.[1]
Review form and statistical synthesis answer different questions
| Form | Main purpose or feature | What the label cannot establish |
|---|---|---|
| Systematic review | Explicit methods for a defined research question | Not automatically a sound or complete review[1] |
| Scoping review | Map the extent and character of literature | Not necessarily an intervention-effect estimate[2] |
| Narrative review | Broader discussion not explicitly systematic | Quality still requires appraisal[3] |
| Meta-analysis | Statistically combine suitable study estimates | Pooling cannot repair every study limitation[1] |
A systematic method makes decisions visible
PRISMA 2020 describes reporting of the question, eligibility criteria, information sources, search, selection, appraisal and synthesis. These details allow readers to understand which evidence was sought and included. A flow diagram or completed checklist is not itself proof that the search was comprehensive or decisions were unbiased. PRISMA is a reporting guideline, not a tool for rating the conduct or methodological quality of the review. Its name therefore does not certify every conclusion.[1]
A scoping review can serve a different purpose
The PRISMA-ScR framework concerns synthesizing evidence and assessing the scope of literature on a topic. Such work can map concepts, evidence types and gaps, or help decide whether a more focused systematic review is warranted. Its reporting framework includes search, selection, data charting and synthesis, with critical-appraisal items when appraisal is performed. A broad map and an estimate of treatment benefit are different products. The review’s question and methods matter more than assuming that every scoping review answers a clinical-effectiveness question.[2]
Narrative synthesis and a narrative review are not interchangeable labels
The SANRA authors describe narrative reviews as literature summaries that are not explicitly systematic and may address a topic broadly. They propose appraisal of aims, searching, referencing, evidence and endpoint data. This is different from a systematic review that uses a transparent narrative or other synthesis because statistical pooling is unsuitable. A useful broad explanation can still have limitations in selection and completeness. The article type alone does not establish either that it is worthless or that it provides a comprehensive effect estimate.[3]
A systematic review does not have to pool incompatible data
SWiM addresses reporting of intervention-effect syntheses when meta-analysis is unavailable or inappropriate. It asks authors to explain study grouping, chosen synthesis methods, prioritization, heterogeneity and limitations. Different methods can answer different questions: describing the distribution or direction of effects is not the same as estimating a weighted average effect. Transparent presentation remains necessary even without a pooled number. A coherent written conclusion should not hide how studies contributed or which data were missing.[4]
Variation between studies needs more than one statistic
Higgins and Thompson introduced measures of statistical heterogeneity, including I-squared, describing how much variation in study estimates relates to heterogeneity rather than sampling error under the model. It is not the percentage of participants who respond differently or a direct measure of clinical importance. Populations, interventions, outcomes and methods can differ even when a statistic appears small. Conversely, substantial statistical variation does not identify its cause. A heterogeneity label needs study context and uncertainty rather than a universal pooling cutoff.[5]
Appraisal examines critical weaknesses rather than counting checklist points
AMSTAR 2 assesses systematic reviews of intervention research, including randomized and nonrandomized studies. Its authors identify potentially critical domains such as protocol, search, exclusions, study bias, synthesis and missing results. They warn against adding item ratings into one overall numerical score because that can disguise a critical weakness. The relevant domains depend on the review and some judgments require methodological and topic expertise. A review can be large and statistically precise while retaining serious methodological limitations.[6]
A living review is an updating process
A living systematic review combines an initial review with continuing, frequent surveillance for new evidence. Its protocol should explain how searching and integration will occur, including how methods changes are reported. “Living” describes maintenance, not an automatic guarantee of stronger evidence.[9]
Evidence & uncertainty
A pooled average can conceal variation between studies. Riley and colleagues explain that the confidence interval for a random-effects average and a prediction interval concerning effects across similar settings describe different uncertainties. A precise average does not promise the same result everywhere.[8]
Selected-source overview · Independent review pendingPublished questions
Are systematic review and meta-analysis interchangeable terms?
Selected-source overview · Independent review pending. Read the evidence context.
No. PRISMA distinguishes the explicit review process from statistical combination of study estimates. A systematic review can be conducted without a meta-analysis.[1]
Evidence behind this answer
- The PRISMA 2020 statement: an updated guideline for reporting systematic reviews
original reporting statement · 2021
Selected original full-text methods, results and limitations inspected; no independent systematic appraisal · Systematic review/meta-analysis definitions, transparency and reporting-versus-conduct distinction
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Does a scoping review necessarily estimate an intervention’s health effect?
Selected-source overview · Independent review pending. Read the evidence context.
No. PRISMA-ScR describes mapping the scope of literature and related objectives. The particular question, appraisal and synthesis determine what conclusions are supported.[2]
Evidence behind this answer
- PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation
original scoping reporting statement · 2018
Original indexed abstract and official author-programme explanation/checklist passages inspected; full explanatory study paper not independently inspected · Scope mapping, review purposes, charting/synthesis and appraisal-if-performed distinction
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Does the narrative-review label make a broad explanation automatically worthless?
Selected-source overview · Independent review pending. Read the evidence context.
No. SANRA addresses appraisal of such reviews. A useful broad purpose still needs transparent aims, evidence and recognition of selection and completeness limits.[3]
Evidence behind this answer
- SANRA—a scale for the quality assessment of narrative review articles
original appraisal tool development and field test · 2019
Selected original full-text methods, results and limitations inspected; no independent systematic appraisal · Narrative definition/purpose; six domains, three editors/30 manuscripts and validation limits
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Does absence of pooling mean a systematic review has failed?
Selected-source overview · Independent review pending. Read the evidence context.
No. SWiM addresses circumstances where pooling is unsuitable or unavailable. The alternative synthesis should explain its methods, contribution of studies and limitations.[4]
Evidence behind this answer
- Synthesis without meta-analysis (SWiM) in systematic reviews: reporting guideline
original reporting method statement · 2020
Selected original full-text methods, results and limitations inspected; no independent systematic appraisal · Alternative synthesis methods, grouping, prioritization, heterogeneity and the question answered
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Does I-squared reveal which patients benefit or supply a universal pooling rule?
Selected-source overview · Independent review pending. Read the evidence context.
No. It describes statistical variation under a model. It does not identify individual responses, the cause of variation or clinical importance by itself.[5]
Evidence behind this answer
- Quantifying heterogeneity in a meta-analysis
original statistical methods paper · 2002
Original indexed abstract and selected original indexed passages inspected; direct original page unavailable or challenge blocked; full text not independently inspected · I-squared/variation versus sampling error; study-context interpretation without fixed cutoff
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Should AMSTAR 2 items be summed into one overall quality score?
Selected-source overview · Independent review pending. Read the evidence context.
Its authors advise against this. Summation can hide critical weaknesses; appraisal depends on domains and the review’s question and methods.[6]
Evidence behind this answer
- AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both
original review appraisal instrument · 2017
Selected original full-text methods, results and limitations inspected; no independent systematic appraisal · Critical domains, context/expertise and warning against a summed overall score
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Are matching conclusions from two overlapping reviews independent replications?
Selected-source overview · Independent review pending. Read the evidence context.
No. Lunny and colleagues show why shared primary-study results require identification and careful synthesis. Repeated counting can create misleading precision.[7]
Evidence behind this answer
- Managing overlap of primary study results across systematic reviews: practical considerations for authors of overviews of reviews
original methods and case studies · 2021
Selected original full-text methods, results and limitations inspected; no independent systematic appraisal · Repeated primary results, overstated precision, citation matrices and six worked cases
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Does a precise random-effects average guarantee the same effect in another setting?
Selected-source overview · Independent review pending. Read the evidence context.
No. The average and between-study variation answer different questions. Model assumptions and similarity to the represented studies remain relevant.[8]
Evidence behind this answer
- Interpretation of random effects meta-analyses
original statistical interpretation paper · 2011
Original indexed abstract and selected original indexed passages inspected; direct original page unavailable or challenge blocked; full text not independently inspected · Average effect versus setting variation, prediction interval and limits of clinical transfer
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Does a living review already include every paper published today?
Selected-source overview · Independent review pending. Read the evidence context.
No. Continuing surveillance still involves searching, assessment and integration. The review should make its timing and any deferred updates visible; the label alone does not establish complete current coverage.[9]
Evidence behind this answer
- Living systematic review: 1. Introduction—the why, what, when, and how
original review methods framework · 2017
University-hosted original published PDF retrieved. Selected definition, protocol, search/update scenarios and funding read. No universal monthly schedule or current publisher policy inferred from the 2017 framework. · Published8-page PDF: introduction/definition, protocol, search and four update scenarios; funding
This answer summarizes selected sources. An independent appraisal of the complete evidence has not been completed.
Random-effects pooling supplies an average under assumptions
Riley and colleagues explain that a random-effects summary estimates an average across study effects rather than one effect shared identically by every setting. A prediction interval can express between-study variation alongside uncertainty in the average. Both depend on the model and available studies; neither guarantees an individual outcome. If studies address importantly different questions, an average may have limited meaning even when mathematically computable. The statistical summary should remain connected to the actual populations, comparisons, follow-up and endpoints represented.[8]
Appraisal tools themselves have limited validation evidence
The SANRA field test involved three editors rating 30 narrative manuscripts submitted to one journal. Agreement was imperfect, and the investigators noted the small sample, developer-raters and need for further validity testing. This is evidence about an appraisal instrument in that setting, rather than proof that using the scale improves health outcomes or ranks every review across journals. A tool can structure scrutiny while still requiring training, judgment and awareness of its own evidence limits.[3]
Search, assessment and integration take time
Finding a new report does not instantly put it in the synthesis. It must be assessed, and incorporation may be deferred with a visible explanation. Check the last search, the version and how newly found evidence was handled. The original methods framework did not define one universal update interval.[9]
References
- The PRISMA 2020 statement: an updated guideline for reporting systematic reviews
Original reporting statement
Page and colleagues · Content/publication date: 2021
Accessed: 2026-10-14 UTC · Selected original full-text methods, results and limitations inspected; no independent systematic appraisal
Cited sections: Systematic review/meta-analysis definitions, transparency and reporting-versus-conduct distinctionOriginal co-publication in Systematic Reviews inspected after BMJ 403/PMC challenge. BMJ n71, PLoS pmed.1003583, Springer s13643-021-01626-4 and other co-publications/PMCs are one underlying statement, not new source allowances. No claim that checklist use completes appraisal.
- PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation
Original scoping reporting statement
Tricco and colleagues; PRISMA Executive · Content/publication date: 2018
Accessed: 2026-10-14 UTC · Original indexed abstract and official author-programme explanation/checklist passages inspected; full explanatory study paper not independently inspected
Cited sections: Scope mapping, review purposes, charting/synthesis and appraisal-if-performed distinctionDirect PubMed returned a shell; official PRISMA scoping page retains the 2018 framework. Year precision retained separately from journal issue October 2. DOI 10.7326/M18-0850 and programme/checklist routes share one statement allowance. A 2026 update-development scoping-review search result was observed, not treated as an already adopted new checklist.
- SANRA—a scale for the quality assessment of narrative review articles
Original appraisal tool development and field test
Baethge, Goldbeck-Wood and Mertens · Content/publication date: 2019
Accessed: 2026-10-14 UTC · Selected original full-text methods, results and limitations inspected; no independent systematic appraisal
Cited sections: Narrative definition/purpose; six domains, three editors/30 manuscripts and validation limitsSelected original background, methods, results and limitations inspected. Developer-raters and one-journal sample are explicit; not a randomized test of health benefits from review appraisal. PMID 30962953/PMC6434870/DOI are one paper.
- Synthesis without meta-analysis (SWiM) in systematic reviews: reporting guideline
Original reporting method statement
Campbell and colleagues · Content/publication date: 2020
Accessed: 2026-10-14 UTC · Selected original full-text methods, results and limitations inspected; no independent systematic appraisal
Cited sections: Alternative synthesis methods, grouping, prioritization, heterogeneity and the question answeredSelected original PMC7190266 methods/guidance inspected after publisher error. SWiM paper is distinct from the Handbook it references; no Handbook text is reproduced or counted as a separate allowance.
- Quantifying heterogeneity in a meta-analysis
Original statistical methods paper
Higgins and Thompson · Content/publication date: 2002
Accessed: 2026-10-14 UTC · Original indexed abstract and selected original indexed passages inspected; direct original page unavailable or challenge blocked; full text not independently inspected
Cited sections: I-squared/variation versus sampling error; study-context interpretation without fixed cutoffOriginal indexed full abstract inspected; direct PubMed shell and publisher full text not independently inspected. Year precision retained; June 15 issue is distinct from any earlier online date not established here. DOI 10.1002/sim.1186/PMID 12111919 are one paper.
- AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both
Original review appraisal instrument
Shea and colleagues · Content/publication date: 2017
Accessed: 2026-10-14 UTC · Selected original full-text methods, results and limitations inspected; no independent systematic appraisal
Cited sections: Critical domains, context/expertise and warning against a summed overall scoreSelected original PMC5833365 development, appraisal guidance and discussion inspected after publisher error. Appraisal of review methods is distinct from certainty by outcome; no Atlas formal appraisal claimed.
- Managing overlap of primary study results across systematic reviews: practical considerations for authors of overviews of reviews
Original methods and case studies
Lunny, Pieper, Thabet and Kanji · Content/publication date: 2021
Accessed: 2026-10-14 UTC · Selected original full-text methods, results and limitations inspected; no independent systematic appraisal
Cited sections: Repeated primary results, overstated precision, citation matrices and six worked casesSelected original methods and cases inspected; PubMed 34233615 direct attempt was 429. PMID 34233615/PMC8265144/DOI are one paper. No universal overlap correction or extra independent evidence from a duplicate review is assumed.
- Interpretation of random effects meta-analyses
Original statistical interpretation paper
Riley, Higgins and Deeks · Content/publication date: 2011
Accessed: 2026-10-14 UTC · Original indexed abstract and selected original indexed passages inspected; direct original page unavailable or challenge blocked; full text not independently inspected
Cited sections: Average effect versus setting variation, prediction interval and limits of clinical transferOriginal indexed article introduction/summary and author-university bibliographic record inspected; direct BMJ failed and full text not independently appraised. Prediction-interval interpretation is bounded by the model and study context, not an individual forecast. PMID 21310794/DOI are one work.
- Living systematic review: 1. Introduction—the why, what, when, and how
Original review methods framework
Elliott JH and colleagues · Content/publication date: 2017
Accessed: 2026-10-20 UTC · University-hosted original published PDF retrieved. Selected definition, protocol, search/update scenarios and funding read. No universal monthly schedule or current publisher policy inferred from the 2017 framework.
Cited sections: Published8-page PDF: introduction/definition, protocol, search and four update scenarios; fundingMethods framework; maintenance is not a quality guarantee.
Scope & review
- Population
- Readers of health reviews; intervention-effect reviews, scoping/narrative forms and original methodological examples
- Exposure or intervention
- Review purposes and explicit methods; statistical synthesis, heterogeneity, overlapping study results and appraisal
- Comparison
- Different review/synthesis methods and appraisal domains; selected tool field tests rather than a health-intervention trial
- Outcomes
- Completeness/transparency, effect estimates, between-study variation, dependence, methodological weaknesses and appraisal reliability
- Timeframe
- Source-specific methodological publications from 2002–2021; source/checklist versions and October retrieval recorded separately
- Jurisdiction
- General evidence interpretation; study populations and methods determine transfer. No clinical effectiveness ranking or completed Atlas systematic review/GRADE rating
Selected-source educational explanation. Official guidance, original study findings and editorial interpretation are distinguished. No exhaustive search, independent systematic synthesis or qualified medical review has been completed.
Sources checked on their individual access dates. Selected sources were checked; ongoing surveillance and qualified review have separate statuses on the coverage page.
Revision history
- 2026-10-14 UTC: Added a bounded evidence-interpretation explainer; independent clinical review and comprehensive synthesis remain pending.
- 2026-10-20 UTC: Added supported foundational, practical and evidence explanations after reading full target bodies and relevant current/pending adjacent coverage. Original sections, questions and sources preserved as exact prefixes; qualified review pending.
What this article still leaves open
- How well do available reviews answer the exact population, comparison and outcome of interest?
- Which original studies and cohorts overlap across reviews and follow-up reports?
- How do synthesis choices and critical methodological weaknesses change interpretation of the apparent average?
Suggest a correction
Submit a content correction to the editorial queue, or download a local note. Leave out personal medical details. Submitted text is stored privately with a browser receipt session and is visible to this browser and the editor. Clearing cookies removes access to your submission history. This is not an account or a medical service.
Your submitted corrections for this article
Explore the topic map
Browse the health domains connected to this article. The map also identifies topics whose articles remain unwritten.