A narrative review is a traditional form of literature review in which an author surveys the published work on a topic and synthesises it into a connected account, using expert judgement rather than a pre-registered, exhaustive protocol. It has no formal eligibility criteria, no requirement to search every database, and no obligation to appraise each source with a validated tool. Its strength is breadth and interpretation: it can place a field in context, trace how ideas developed, and frame open questions. Its weakness is that the selection of sources is, by design, the author’s, which leaves room for selective citation.
How a narrative review differs from a systematic review
The defining contrast is method. A systematic review built on a protocol pre-specifies its question, its search strategy, its inclusion and exclusion criteria, and its appraisal method before any results are seen, precisely so that another team could reproduce it and reach the same set of included studies. A narrative review does none of this as a requirement. The author chooses what to read and what to feature, which makes the process fast and flexible but not reproducible and vulnerable to selection bias and citation bias, the tendency to cite work that supports a prior view. The fuller comparison, including when each is the right instrument, is set out in our piece on the systematic review versus the literature review. In short: a systematic review aims to remove the author from the selection; a narrative review relies on the author’s expertise to make it.
When a narrative review is the right choice
Narrative reviews are not a lesser systematic review; they answer different questions. They are well suited to broad topics where a single answerable question would be artificial, to introducing or teaching a field to newcomers, to theoretical or conceptual synthesis that draws threads across disparate literatures, and to agenda-setting pieces that argue for new directions. A clinician wanting to understand the landscape of a disease, or a researcher framing the introduction to a thesis, is usually better served by a narrative review than by a narrowly scoped systematic one. When the question is a focused, answerable comparison, such as whether one treatment beats another on a defined outcome, the narrative form is the wrong tool and a systematic review or even a quantitative pooling of the evidence is needed instead. Choosing the form to fit the question is the first decision, and the wider menu of options sits in our overview of the main types of literature review.
How to make a narrative review rigorous
A narrative review need not be unsystematic in spirit, and the best ones borrow discipline from their systematic cousins without pretending to be one. Several practices raise the standard markedly. First, state the scope and rationale: tell the reader what the review covers, what it deliberately excludes, and why, so the boundaries are visible. Second, describe how sources were found, naming the databases searched and the rough date range and terms, even if the search was not exhaustive, so readers can judge its coverage. Third, organise thematically rather than paper by paper, so the review synthesises ideas instead of listing summaries. Fourth, acknowledge contradictory evidence explicitly rather than citing only the supportive studies. Fifth, declare conflicts of interest and the author’s standpoint, since interpretation is the method. A useful halfway house is to be transparent about searching in the same spirit as the planned stages of a structured review, even where full reproducibility is not the goal. Done this way, a narrative review is honest about what it is: an informed interpretation, clearly bounded, rather than a comprehensive verdict.
Narrative review versus narrative synthesis
These two terms are routinely confused, and the difference matters. A narrative review is a whole review type, defined by its non-systematic, expert-led method. A narrative synthesis is something else entirely: it is a structured, textual method of combining findings inside a systematic review when the included studies are too clinically or methodologically diverse to pool statistically. In a narrative synthesis the search, selection, and appraisal are still fully systematic and reproducible; only the final combination step is descriptive rather than a meta-analysis, following a recognised framework with steps such as grouping studies, tabulating effects, and exploring patterns. The contrast between pooling numbers and synthesising in words is laid out in our guide to narrative synthesis against meta-analysis. The single-word overlap hides a real distinction: a narrative review relaxes the whole method, whereas a narrative synthesis relaxes only the statistical combination within an otherwise rigorous systematic review.
Limitations and common mistakes
The chief limitation of a narrative review is the one built into its method: because selection is not systematic, the conclusions can reflect the author’s prior beliefs as much as the evidence, and there is no audit trail to check. It cannot, and should not, be used to make a definitive claim about the size of an effect, since it neither searched exhaustively nor appraised each study’s risk of systematic error. The recurring mistakes are equally predictable. The first is dressing a narrative review up as a systematic one, using the words “systematic” or even “PRISMA” without any of the underlying method, which misleads readers and editors. The second is cherry-picking supportive studies and ignoring the inconvenient ones. The third is listing rather than synthesising, producing a string of paper summaries with no connective argument. The fourth is making strong causal or quantitative claims the method cannot support. A narrative review that is honest about its scope, open about how it found its sources, balanced in its coverage, and modest in its conclusions remains one of the most valuable forms of academic writing; problems arise only when it borrows the authority of a method it did not use.