When you want to know what research says about a question, a review article is the natural place to start. But "review" covers two quite different things. A narrative review is an expert's guided tour of a topic. A systematic review is closer to a study in its own right, with methods designed so that someone else could repeat it and get the same set of studies.
What a narrative review does well
A narrative review is written by someone who knows a field. It can explain history, connect ideas across areas, highlight promising directions and give readers a map. It is often the best way to get oriented.
Its weakness is selection. The author chooses which studies to discuss, and there's usually no record of what was searched for or why some papers were left out. Two experts can write narrative reviews on the same question and reach different conclusions, each citing different evidence.
What makes a review systematic
A systematic review tries to remove that selection problem by fixing the rules in advance:
- A focused question. Often framed as population, intervention or exposure, comparison and outcome.
- A pre-specified protocol. The plan is written down before the search, and often registered publicly (for example in PROSPERO).
- A comprehensive search. Several databases, with the full search strings reported, plus attempts to find unpublished studies.
- Explicit inclusion criteria, applied to each study, usually by two reviewers independently.
- Risk-of-bias assessment. Each included study is rated on how much its design and conduct could distort its result.
- A transparent synthesis. Sometimes a meta-analysis that pools numbers, sometimes a structured narrative synthesis when the studies are too different to combine.
- A flow diagram (such as PRISMA) showing how many records were found, screened, excluded and included.
The goal is not to remove judgment but to make every judgment visible.
Systematic does not mean definitive
A systematic review can only be as good as the studies it includes. If those studies are small, biased, or selectively published, a careful review will faithfully summarize weak evidence. Reviews also age: a question that was settled five years ago may look different after new trials.
How to tell which one you're reading
- Is there a methods section describing databases, search terms and dates?
- Are inclusion and exclusion criteria stated?
- Is there a flow diagram and a table of included studies?
- Was the quality or risk of bias of each study assessed?
If the answers are mostly no, you are reading a narrative review. That's fine, as long as you treat it as an informed perspective rather than a complete account of the evidence.
More research explainers: https://www.crushsci.com