What a narrative review is
A narrative review, sometimes called a traditional or literature review, summarizes and discusses what has been written on a topic. The author chooses the sources, interprets them and builds an argument or a framework. Most textbook chapters, many invited journal articles and many "state of the field" papers are narrative reviews.
The format has real uses. It can explain a concept to a general audience, trace the history of an idea, bring together findings from different disciplines, discuss mechanisms and clinical reasoning that numbers alone cannot capture, and point to gaps. In emerging fields where few primary studies exist, or where the studies are too diverse to compare, a narrative review may be the only feasible form.
Grant and Booth's typology of reviews lists the narrative review among the types that do not require a systematic search, formal appraisal or protocol. The label does not mean the work is poor. It means the method is not fixed, so the reader has to rely on the author's judgment, and the author carries the burden of making that judgment visible.
The risk of bias in narrative reviews
The main concern is selection. If the author does not state how studies were found and chosen, a reader cannot know whether the cited studies represent the evidence or only support one view. Mulrow's 1987 assessment of medical review articles found that most did not use specific methods to identify, select and validate information. Later work, including Antman and colleagues' comparison in 1992, showed that expert narrative reviews could lag behind cumulative meta-analysis of trials or recommend treatments that the pooled evidence did not support.
Other risks follow from the same cause. Authors may cite the studies they know, which favors their own group's work and well known journals. They may pay more attention to positive, statistically significant findings. They may treat studies of different sizes and designs as equal in weight, or let vivid case reports outweigh large trials. None of these problems is inevitable, but all are easier without a stated method.
Because of these risks, narrative reviews should not be used to estimate the size of an effect, to settle a disputed clinical question or to produce practice recommendations. When those are the aims, a systematic review is the suitable tool.
Making a narrative review more transparent
A narrative review can be made more reliable without turning it into a systematic review. Green and colleagues' guidance for authors, and the SANRA scale developed by Baethge and colleagues for editors and readers, point to similar practices.
| Element | What to do |
|---|---|
| Justification | Explain why the review is important for readers and what is not already covered |
| Aims | State the aims and the scope; a vague topic leads to a vague review |
| Literature search | Describe the sources, the dates and the search terms, even if the search is not exhaustive |
| Referencing | Cite key evidence accurately and include evidence that goes against the author's view |
| Evidence presentation | Distinguish study design and size; say what each study can and cannot show |
| Data presentation | Use tables or figures where they make the evidence clearer |
SANRA (Scale for the Assessment of Narrative Review Articles) has six items, each scored from 0 to 2, covering these same domains: the justification, the statement of aims, the description of the literature search, referencing, scientific reasoning and appropriate presentation of data. It was designed as a short tool for editors, reviewers and authors. Its scores have no validated cut-off for "good" or "poor", and it should be used to structure feedback and not to rank articles.
Searching without being exhaustive
A narrative review does not need to find every study, but it should be able to say what was searched. A short methods paragraph is enough: the databases, the date of the search, the main terms, the languages and any limits, and how additional sources were found, for instance through reference lists or expert suggestions. A search of at least one major database, such as PubMed, with a recorded strategy is a reasonable minimum for a biomedical topic.
The author should state how studies were chosen to be discussed. Perhaps the review prioritizes systematic reviews, then randomized trials, then other designs; perhaps it prioritizes recent work. Whatever the rule, saying so lets the reader judge. The review should also address evidence that contradicts the author's argument. A section on contrary findings and uncertainty makes a narrative review more credible.
Structure and style
A narrative review needs a clear structure, even without a protocol. The introduction explains the question and why it matters. The body is organized by theme, by time, by mechanism or by population, whichever helps the reader follow the argument, and uses headings that say what each section contributes. A synthesis section states what the evidence shows as a whole, where it is weak and what research would resolve the main uncertainties.
Tables and figures can carry a lot of weight. A table of key studies, with design, population, size and main finding, shows the base of evidence at a glance. A conceptual figure can show a proposed framework. Statements should be linked to specific sources, with the strength of the evidence made clear: "a randomized trial of 400 patients found" tells the reader more than "research shows".
Language should match the evidence. Words such as "proves" or "establishes" are rarely justified. If an effect was seen in one small study, say so. The conclusions should follow from the evidence presented, and speculation should be marked as such.
Reading a narrative review critically
Readers, peer reviewers and editors can ask a short set of questions of any narrative review. Is the purpose clear, and does the review stay within it? Is there a methods statement, and does it say how sources were found? Are the key studies cited accurately, and does the text match what those studies found? A common fault is citing a review for a claim that the review itself attributes to another source, so that errors are passed along unchecked. Where possible, check the main claims against the primary studies.
Next, look for balance. Does the review discuss evidence that cuts against its thesis? Does it distinguish randomized from observational evidence, and large studies from small ones? Are authors' own earlier papers cited in proportion to their weight in the field? Finally, look at the conclusions. Do they follow from the evidence shown, or do they go beyond it into recommendations that the cited studies cannot support? A review that claims to settle a question without a systematic method deserves caution.
Editors may ask authors to add a methods paragraph, to temper strong claims and to state limits. These requests improve the paper and are not a sign that the review is rejected. Authors can anticipate them by writing the methods paragraph at the start.
Upgrading a narrative review
If a narrative review turns out to address a question that others will want to answer reliably, it can be upgraded. The usual steps are to write a protocol, to turn the informal search into a structured one with at least two databases, to screen records against written criteria, to extract data into a standard form and to appraise study quality. The earlier narrative work is not wasted: it supplies the background, the key terms and a first list of studies. Upgrading changes the claim the review can make, because a documented method lets the authors say how complete the evidence is and how certain the conclusions are.
Choosing between a narrative and a systematic review
The choice follows the purpose. If the aim is to teach, to frame a debate, to describe a field or to explore ideas, a narrative review is appropriate. If the aim is to answer a specific question about effects, accuracy or prevalence in a way others can reproduce, a systematic review is better. If the aim is to map the extent of research without judging its quality, a scoping review fits. If the topic needs many designs brought into one account, an integrative review may fit.
Some journals publish narrative reviews only on invitation. Others accept submissions but ask for a methods statement. Check the author guidelines. A well constructed narrative review with a stated search can be more useful to readers than a poorly conducted systematic review, and the label alone is not a measure of quality.
When a narrative review is planned as a first step, its search and tables can feed later work. Many systematic reviews begin with a narrative overview to refine the question and to test whether a systematic review is already available.
Limitations
The narrative review is open to selection bias, depends on the author's judgment and cannot be reproduced exactly. Without appraisal of study quality, strong and weak studies may be given equal weight. It does not give an effect size or a measure of certainty, and conclusions drawn from it are weaker than those from a systematic review. SANRA helps to structure assessment but has no validated threshold. Readers should treat the conclusions as an informed interpretation and look for a stated method and for acknowledgment of contrary evidence.
Evidence synthesis informs decisions about groups and policy. It does not replace professional judgment about an individual.
How we can help
Support for a narrative review
The work can be supported at different depths. Choose what you need, and the scope is agreed in writing before work begins.
Protocol and registration
A protocol and analysis plan written before the search, with registration prepared where appropriate.
Search and screening
Documented search strategies, screening and the flow diagram.
Appraisal and synthesis
Data extraction, risk of bias, certainty of evidence and the synthesis.
Manuscript and submission
Optional: the full paper, the reporting checklist and the submission materials.
Frequently asked questions
What is a narrative review?
An overview of the literature on a topic written by an author or group, without a fixed protocol for searching, selecting or appraising studies.
Is a narrative review less valuable than a systematic review?
It answers different questions. It suits background and framing, but it is more open to bias and should not be used to estimate effects or settle clinical questions.
Does a narrative review need a search?
It does not need an exhaustive one, but it should say what was searched, when and how studies were chosen, so readers can judge selection.
What is SANRA?
The Scale for the Assessment of Narrative Review Articles, a six-item tool covering justification, aims, search description, referencing, reasoning and data presentation.
Can I register a narrative review?
Most registries do not accept them, since there is no fixed protocol. You can still publish a methods description with the review.
When should I choose a systematic review instead?
When you need a reproducible answer to a defined question, such as the effect of an intervention or the accuracy of a test.
References
- Green BN, Johnson CD, Adams A. Writing narrative literature reviews for peer-reviewed journals: secrets of the trade. J Chiropr Med. 2006;5(3):101-117.
- Baethge C, Goldbeck-Wood S, Mertens S. SANRA - a scale for the quality assessment of narrative review articles. Res Integr Peer Rev. 2019;4:5.
- Mulrow CD. The medical review article: state of the science. Ann Intern Med. 1987;106(3):485-488.
- Antman EM, Lau J, Kupelnick B, Mosteller F, Chalmers TC. A comparison of results of meta-analyses of randomized control trials and recommendations of clinical experts. JAMA. 1992;268(2):240-248.
- Grant MJ, Booth A. A typology of reviews: an analysis of 14 review types and associated methodologies. Health Info Libr J. 2009;26(2):91-108.
- Gasparyan AY, Ayvazyan L, Blackmore H, Kitas GD. Writing a narrative biomedical review: considerations for authors, peer reviewers, and editors. Rheumatol Int. 2011;31(11):1409-1417.