Is an umbrella review the same as a meta-analysis of meta-analyses?

Related but not identical. A meta-analysis of meta-analyses pools effect estimates extracted from reviews — methodologically difficult because of overlap. Umbrella reviews more often use narrative synthesis and quality weighting. Both are 'reviews of reviews' families; check authors' methods carefully.

Can I do an umbrella review for my dissertation?

A full umbrella review is usually too large for a solo UK medical student in one semester. Appraising an existing umbrella review, or conducting a bounded comparison of two or three systematic reviews, is more realistic. Discuss scope with your supervisor early.

Which quality tool do I use on included reviews?

AMSTAR 2 for methodological quality of each included systematic review. ROBIS when you need to assess risk of bias in how conclusions were drawn. Do not apply ROB 2 to the reviews — ROB 2 is for randomised trials within them.

What is corrected covered area (CCA)?

CCA measures how much the same primary studies appear across multiple included reviews. High overlap means evidence is not independent across reviews. Umbrella reviews should report CCA or an equivalent overlap metric.

How do I search for systematic reviews only?

Use filters for systematic reviews in MEDLINE and Embase, search Epistemonikos, and check Cochrane Library. Pre-specify whether scoping reviews or rapid reviews are eligible. Document search strings in an appendix.

Strata Academy

Umbrella Reviews & Overviews of Systematic Reviews Explained

How reviews of reviews differ from primary systematic reviews, AMSTAR 2 at scale, and when umbrella methods suit student appraisal coursework

Quick answer

An umbrella review (overview of systematic reviews) synthesises evidence from multiple existing systematic reviews on a related topic — not primary studies. You appraise included reviews with AMSTAR 2 (and often ROBIS), assess overlap of primary studies, and may re-interpret GRADE certainty across reviews. Full umbrella reviews are ambitious for solo student projects; appraising one published umbrella review is excellent journal club material.

1. What is an umbrella review?

An umbrella review — also called an overview of systematic reviews or review of reviews — collects and synthesises findings from multiple systematic reviews (and sometimes meta-analyses) addressing related questions in a broad clinical area.

The unit of inclusion is the systematic review, not the individual trial. Primary studies may appear in several included reviews; overlap distorts naive pooling if ignored.

Umbrella reviews suit questions where many systematic reviews already exist — diabetes management, depression treatments, COVID-19 therapeutics — and decision-makers need a map of what reviews agree on, where they conflict, and what GRADE certainty looks like across the field.

For students, reading an umbrella review is a fast way into a topic. Conducting one requires systematic search for reviews, dual screening, AMSTAR 2 on every included review, overlap analysis, and often qualitative evidence mapping — heavier than a focused primary systematic review on one PICO.

2. Umbrella review vs standard systematic review

A standard systematic review searches for primary studies (RCTs, cohorts, etc.), applies study-level risk of bias (ROB 2, ROBINS-I), and may meta-analyse trial-level data.

An umbrella review searches for systematic reviews, appraises review-level quality (AMSTAR 2, ROBIS), extracts review conclusions and GRADE ratings, and may compare overlapping primary studies across reviews.

PRISMA reporting applies to both, but umbrella methods extensions emphasise documenting how reviews were selected, how overlap was measured (e.g. corrected covered area), and how conflicting review conclusions were handled.

Students sometimes propose 'umbrella reviews' when they mean a broad primary review. Supervisors will redirect if your search targets trials, not existing reviews.

3. Core methodological steps

Define eligibility: which systematic reviews count (Cochrane only? meta-analyses required? language limits?). Pre-register on PROSPERO or OSF — same transparency expectations as primary reviews.

Search comprehensively for reviews: databases (MEDLINE, Embase, Epistemonikos), hand-searching guideline appendices, citation chasing. Grey literature may matter if policy reviews sit in reports.

Dual independent screening and selection. Extract: review PICO, number of primary studies, AMSTAR 2 score, GRADE conclusions, overlap with other included reviews.

Assess overlap of primary studies across included reviews using a validated metric (e.g. corrected covered area — CCA). High overlap means conclusions are not independent — interpret accordingly.

Synthesise narratively or in tables mapping review conclusions per outcome. Re-pooling trial data from reviews without accessing individual patient data is generally discouraged unless methods are explicit and overlap is low.

PROSPERO registration with 'overview of reviews' eligibility
AMSTAR 2 for each included review — critical flaws downgrade trust
ROBIS when assessing bias in review conclusions
Overlap matrix — which primary studies appear in which reviews
Conflict resolution when reviews disagree on the same outcome

4. AMSTAR 2 in umbrella reviews

Every included systematic review should receive AMSTAR 2 assessment. Umbrella conclusions are only as strong as the weakest critical-flaw reviews you treat as evidence.

Common pattern: one high-quality Cochrane review and several low-quality narrative 'systematic' reviews with PubMed-only searches. Umbrella authors must not average conclusions without weighting by methodological quality.

Students appraising an umbrella review for coursework should spot-check AMSTAR 2 ratings on two or three included reviews against your own reading — authors sometimes overstate included review quality.

Critical AMSTAR 2 flaws (no comprehensive search, no duplicate screening, no RoB assessment) mean that review's conclusions should be downweighted or excluded from synthesis.

5. Overlap and double-counting

The same landmark RCT often appears in five systematic reviews on slightly different diabetes questions. If an umbrella review treats each review's positive conclusion as independent evidence, it double-counts the same trial five times.

Corrected covered area (CCA) quantifies overlap: low CCA suggests reviews are distinct; high CCA means synthesis is effectively repeating the same primary evidence.

When overlap is high, umbrella authors should emphasise the highest-quality review (often Cochrane) rather than vote-counting across redundant reviews.

In journal club, ask: Did the umbrella authors report overlap? Would conclusions change if only AMSTAR 2 high-confidence reviews were included?

6. GRADE and certainty across reviews

Included systematic reviews may each present GRADE certainty ratings for the same outcome. Umbrella reviews should compare these — do Cochrane and non-Cochrane reviews agree on certainty? If not, explain why (different included trials, different RoB handling, different effect sizes).

Umbrella authors rarely re-run formal GRADE from scratch unless they access trial-level data. More often they map certainty statements across reviews in a summary of findings table.

Students should read umbrella conclusions cautiously when all included reviews are 'low' AMSTAR 2 — the umbrella's GRADE synthesis may inherit compounded uncertainty.

7. What students can realistically do

Appraise one published umbrella review for SSC or journal club: apply AMSTAR 2 to the umbrella itself, spot-check two included reviews, assess overlap reporting, and critique conflict resolution.

Propose a 'mini-overview' only with supervisor support: e.g. compare three Cochrane reviews on one drug class — bounded scope, narrative comparison, no claim to full umbrella methods.

Do not promise a field-wide umbrella review in a 12-week dissertation unless you have institutional team support and a very narrow clinical area.

Use StrataResearch to appraise the umbrella review PDF with AMSTAR 2 and PRISMA-aligned feedback, then manually verify AMSTAR 2 on one included review.

Journal club on a published umbrella review — highly recommended
Mini-overview of 2–3 high-quality reviews — possible with tight scope
Full umbrella review solo — rarely feasible
Compare umbrella conclusions to latest primary trials — are overviews outdated?

Interactive version (quizzes, walkthroughs) loads when JavaScript is enabled.