Contents
On this page
- The traditional hierarchy and the apex position of systematic reviews
- The documented critique of ranking by design type alone
- GRADE: rating the certainty of a body of evidence for each outcome
- Application to a specific body of evidence
- The position of well-conducted observational evidence
- Teaching and presenting the pyramid without overstatement
Academy · Research Writing · Levels of Evidence Pyramid
Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach
The traditional evidence pyramid ranks study designs in a fixed hierarchy with systematic reviews at the apex, but a widely cited 2016 paper in BMJ Evidence-Based Medicine argues that the pyramid is outdated as a sole ranking tool. GRADE assesses the certainty of a body of evidence for a specific outcome. Study design contributes to the assessment, alongside factors that can lower or raise certainty. This article examines the hierarchy, the critique and the GRADE approach.
The levels-of-evidence pyramid is a visual hierarchy used to rank study designs by their typical susceptibility to bias, with systematic reviews and meta-analyses of randomized controlled trials traditionally placed at the apex, individual randomized trials beneath them, then cohort studies, case-control studies, case series, and expert opinion at the base. The pyramid remains a useful teaching device for explaining why some designs are generally more reliable than others, but its rigidity, ranking by design type alone, regardless of how well or poorly any individual study of that type was actually conducted, has drawn sustained methodological criticism, most prominently from the GRADE working group's alternative framework.
Learning objectives
- The ranking of the traditional pyramid, and specifically why systematic reviews and meta-analyses occupy its apex
- The documented critique that the pyramid ranks design type without accounting for the quality of execution
- How the GRADE approach assesses evidence quality explicitly rather than by a fixed hierarchy
Common misconceptions
- “A systematic review or meta-analysis is necessarily higher-quality evidence than any individual randomised trial by virtue of its position on the pyramid.” Hatem Mostafa Murad and colleagues’ 2016 paper in BMJ Evidence-Based Medicine, ‘New evidence pyramid’, argues explicitly that a poorly conducted systematic review of flawed trials can be less trustworthy than a single large, rigorously conducted randomised trial, and proposes the explicit quality rating of GRADE as a necessary complement to, rather than a replacement for, the approximate ordering of the pyramid.
The traditional hierarchy and the apex position of systematic reviews
The traditional evidence pyramid orders study designs from most to least susceptible to bias, in descending order: systematic reviews and meta-analyses of randomized controlled trials, individual randomized controlled trials, cohort studies, case-control studies, case series and case reports, and expert opinion or background information at the base1. The rationale for placing systematic reviews and meta-analyses at the apex is that, when well conducted, they combine evidence from multiple randomised trials, reducing the influence of chance findings, sampling variation or idiosyncratic context in any single study, and producing a pooled effect estimate with a narrower confidence interval than any individual contributing trial could achieve alone2. Athanasia Haidich's 2010 methodological overview in Hippokratia describes this pooling function as meta-analysis's central statistical contribution: converting a set of individually underpowered or inconsistent studies into a single estimate with greater statistical power than any one study alone2.
The documented critique of ranking by design type alone
The 2016 paper by Murad and colleagues, published in BMJ Evidence-Based Medicine, argues that the central weakness of the traditional pyramid is that it ranks evidence strictly by study design, irrespective of how rigorously any specific study of that design was conducted or how directly it applies to the clinical question under consideration3. A systematic review built from several small trials at high risk of bias with inconsistent results does not automatically outrank a single, large, well-conducted randomised trial merely because systematic reviews occupy a higher position on the traditional pyramid; the quality of the review depends entirely on the quality of the studies it pooled3. The authors propose a revised visual model that places systematic reviews and meta-analyses beside, rather than strictly above, individual study designs, presenting a systematic review as a lens applied to the underlying evidence rather than as a design category that is inherently superior3.
GRADE: rating the certainty of a body of evidence for each outcome
The 2008 BMJ paper by Gordon Guyatt and colleagues introducing the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach addresses the same weakness from a different perspective: rather than ranking evidence by design category, GRADE explicitly rates the certainty of evidence for each specific outcome, taking study design as an initial marker but then adjusting that rating up or down according to factors including risk of bias, inconsistency across studies, indirectness to the clinical question, imprecision and publication bias, or, in the case of observational studies, a large effect size, a dose-response gradient, or plausible confounding that would reduce rather than create the observed effect4. Under GRADE, a well-conducted observational study showing a large, consistent effect can be rated as higher-certainty evidence than a poorly conducted randomized trial, a conclusion the rigid pyramid cannot represent4. GRADE has since been adopted as the standard approach for grading evidence certainty by the Cochrane Collaboration and by guideline-development bodies including the World Health Organization; in this approach, study design contributes to the assessment alongside factors that can lower or raise certainty5.
Application to a specific body of evidence
Taken together, the traditional pyramid and GRADE serve complementary rather than competing purposes when a specific body of evidence is read. The pyramid remains useful as an initial, approximate orientation, indicating which designs are generally less exposed to confounding and bias before any specific study is examined, which is why it continues to appear in most teaching of evidence-based medicine despite the documented criticism of its use in isolation2. GRADE then performs the specific assessment that the pyramid cannot: it asks, for a particular systematic review or body of trials and a particular outcome, how much risk of bias was present, how consistent the results were across studies, how directly the studied population and intervention correspond to the question, how precise the pooled estimate is, and whether publication bias is a plausible concern given the pattern of available studies4. A practical principle that follows from both documents is never to cite a study’s design type alone as evidence of its quality, since describing a study as a randomised trial or a meta-analysis identifies a category, not a verdict, but instead to ask the GRADE questions explicitly of the specific evidence under consideration, documenting the answers rather than relying on position in the pyramid as a shortcut. This principle is most important where the pyramid and the actual evidence may disagree, as with a design lower on the pyramid that was conducted and reported with unusual rigour, or a design higher on the pyramid that pooled poorly matched or poorly conducted studies.
The position of well-conducted observational evidence
Observational studies, particularly cohort and case-control designs, are placed below randomised trials in the traditional pyramid because, as a category, they are more exposed to confounding, that is, unmeasured differences between the compared groups that could explain an observed association instead of the exposure or intervention under study. The GRADE framework does not dispute this general tendency but specifies the conditions under which it can be overridden for a given body of evidence: a large, consistent effect size that is unlikely to be fully explained by plausible confounders; a dose-response gradient, in which increasing exposure is associated with increasing effect in a pattern difficult to attribute to confounding alone; or a situation in which all plausible confounding would be expected to reduce rather than create the observed effect. Each of these can raise the certainty rating of a body of observational evidence under GRADE4. This is important in practice for questions on which randomised trials are difficult, unethical or not yet conducted, such as long-term harms, rare outcomes or exposures that cannot ethically be randomised, where well-conducted observational evidence may be the best evidence currently available rather than an inferior substitute to be distrusted automatically because of its position on the traditional pyramid. Both frameworks converge on recommending that a position on the pyramid be read as an initial assumption about typical risk of bias, while remaining open to GRADE’s specific, documented conditions for revising that assumption upwards or downwards for the evidence under consideration.
Teaching and presenting the pyramid without overstatement
Because the pyramid remains a useful introductory teaching device, the revised model of Murad and colleagues does not propose discarding it but redrawing it with two specific changes that address its overstatement directly: positioning systematic reviews and meta-analyses as a lens that can be applied across the other design types rather than as a separate, superior tier above them, and giving explicit visual emphasis to the idea that the quality of a study’s execution can move it within, or even beyond, the typical position of its design category3. For those presenting this concept to students, trainees or a non-specialist audience, a straightforward remedy is available even without redrawing the diagram: it should be stated explicitly, alongside the pyramid, that the pyramid describes the typical susceptibility of a design category to bias and does not guarantee the quality of any individual study of that design, and GRADE should be identified as the tool that makes the study-specific judgement the pyramid alone cannot. This two-part presentation, combining the familiar visual with an explicit caveat about what it does and does not claim, preserves the pedagogical value of the pyramid while avoiding the overstatement identified in the critique by Murad and colleagues, and it corresponds to the way in which the Cochrane Handbook recommends introducing study-design hierarchies to new systematic reviewers: as an initial orientation to be followed immediately by the specific, criterion-based assessment that the evidence requires.
Key points
Key points
The traditional evidence pyramid remains a useful teaching device, but its fixed ranking by design type alone has been directly challenged. The explicit and adjustable quality rating of GRADE is now the standard complement, because the actual conduct and relevance of a study, and not merely its design category, determine the degree of confidence it warrants.
Practical application
Practical application
- Position on the pyramid should be treated as a starting point rather than a final judgement. The position of a study on the traditional pyramid indicates the typical risk of bias of its design category, not the actual quality of the specific study under consideration.
- The GRADE domains should be applied explicitly when a body of evidence is assessed. Risk of bias, consistency, directness, precision and publication bias should be assessed for each outcome rather than reliance being placed on design type alone.
- When a systematic review is appraised, the studies it pooled should be examined, and not merely its status as a systematic review. A review of several low-quality trials does not necessarily outrank a single large, well-conducted trial; its credibility depends on the studies it includes.
Frequently asked questions
It is not incorrect, but it is incomplete on its own. The 2016 critique by Murad and colleagues argues that it is outdated as a sole ranking tool because it disregards the quality of execution, and proposes the explicit assessment of GRADE as a necessary complement, rather than arguing that the approximate ordering of the pyramid is reversed.
Yes, under GRADE. A well-conducted observational study with a large, consistent effect can be rated as higher-certainty evidence than a poorly conducted randomized trial, which the fixed pyramid model cannot represent.
Pooling data from multiple studies increases the effective sample size behind the combined estimate, which generally increases statistical power and narrows the confidence interval compared with any single contributing study, as described in Haidich's methodological overview.
The Cochrane Collaboration and guideline-development bodies including the World Health Organization use GRADE as their standard approach for rating the certainty of evidence underlying specific recommendations.
No. GRADE assesses the certainty of a body of evidence for a specific outcome. Study design contributes to the assessment, alongside factors that can lower or raise certainty. Murad and colleagues frame their revised model as placing systematic reviews beside, not above, other designs, and as a complement to GRADE’s criterion-based assessment, rather than discarding the pyramid’s basic logic altogether.
References
- Murad MH, Asi N, Alsawas M, Alahdab F. New evidence pyramid. Evid Based Med. 2016;21(4):125-127. read 10 October 2026. View on BMJ Evidence-Based Medicine
- Haidich AB. Meta-analysis in medical research. Hippokratia. 2010;14(Suppl 1):29-37. read 10 October 2026. View on Hippokratia (PubMed)
- Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926. read 10 October 2026. View on BMJ
- Higgins JPT, Thomas J, Chandler J, et al, eds. Cochrane Handbook for Systematic Reviews of Interventions. read 10 October 2026. View on Cochrane Training
- Murad MH, Asi N, Alsawas M, Alahdab F. New evidence pyramid. Evid Based Med. 2016;21(4):125-127. read 10 October 2026. View on BMJ Evidence-Based Medicine
- Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926. read 10 October 2026. View on BMJ
- Haidich AB. Meta-analysis in medical research. Hippokratia. 2010;14(Suppl 1):29-37. read 10 October 2026. View on Hippokratia (PubMed)
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. read 10 October 2026. View on BMJ
- World Health Organization. WHO Handbook for Guideline Development, 2nd ed. (GRADE methodology). read 10 October 2026. View on WHO
- Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926. read 10 October 2026. View on BMJ
Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
Pubrica Academy. (2026, October 10). Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach. Pubrica. https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/
Pubrica Academy. “Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach.” Pubrica, 10 Oct. 2026, pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/.
Pubrica Academy (2026) Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach. Pubrica. Available at: https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/ (Accessed: 15 October 2026).
Pubrica Academy. “Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach.” Pubrica. October 10, 2026. https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/
Pubrica Academy, ‘Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach’ (Pubrica, 10 October 2026) <https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/> accessed 15 October 2026.
Pubrica Academy. Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach. [Internet]. Pubrica; 2026 Oct 10 [cited 2026 Oct 15]. Available from: https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/
Pubrica Academy, “Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach,” Pubrica, Oct. 10, 2026. [Online]. Available: https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/
@misc{/la_evidence_based_medicine_levels_evidence_pyramid_2026,
author = {{Pubrica Academy}},
title = {Evidence-Based Medicine: The Levels of Evidence Pyramid, the Position of Meta-Analyses and the GRADE Approach},
howpublished = {Pubrica},
year = {2026},
month = {oct},
url = {https://pubrica.com/academy/study-guide/research-writing/evidence-based-medicine-levels-evidence-meta-analyses/},
note = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
GRADE Quality Assessment Quick Sheet
This checklist may be used to rate a body of evidence beyond its position on the pyramid.
Starting point: study design (RCT = high, observational = low) Downgrade for: [ ] risk of bias [ ] inconsistency across studies [ ] indirectness to the question [ ] imprecision [ ] publication bias Upgrade for (observational only): [ ] large effect size [ ] dose-response gradient [ ] plausible confounding would reduce the effect Final certainty rating: High / Moderate / Low / Very low
Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.
How Pubrica can help
Pubrica’s evidence synthesis team applies GRADE alongside the traditional evidence hierarchy when assessing a body of literature, so that conclusions reflect the actual conduct and relevance of studies rather than design category alone.