Systematic review versus narrative review: the pre-specified protocol
A systematic review works to a protocol, a written plan fixed in advance, so that another reviewer can repeat and check its methods.
Remember
A review is systematic because it follows a protocol set in advance, with explicit methods that another reviewer can repeat. Whether the results are combined statistically, which study designs are included and who wrote the review do not make a review systematic.
A systematic review is a study of studies. It sets out to collect all the relevant evidence on one focused enquiry, such as whether a drug lowers blood pressure more than placebo, by following a plan written before the work begins, rather than leaning on whichever papers an author happens to know.
That plan is the pre-specified protocol. It sets out what the review aims to find out, the search strategy (where and how the studies will be found), the eligibility criteria (the rules for which studies are included or excluded) and the analysis plan (how the results will be combined or summarised).
Because the plan exists in advance, the methods are reproducible: another reviewer following the same protocol should reach the same conclusions. Choosing studies by rules rather than by preference also limits bias (a systematic error that pushes results in one direction). The search is meant to be comprehensive, across several databases and grey literature (material produced outside commercial journals, such as reports and theses).
A traditional narrative review summarises the literature in words and does not necessarily follow a plan like this, so a reader cannot easily check how its studies were chosen.
Other things often seen in systematic reviews do not define one. A meta-analysis (a statistical pooling of the results of several studies into one estimate) is optional, and the results can be described in words or in numbers. The included studies need not be randomised controlled trials, meaning original research studies in which investigators randomly allocate participants to treatments.
Key facts
- Defining feature
- a protocol fixed in advance, which makes the methods explicit and repeatable by another reviewer
- Protocol contents
- the focus of the review, the search strategy, the eligibility criteria and the analysis plan
- Search
- a systematic review aims for a comprehensive search, across several databases and grey literature, not a single database
- Meta-analysis
- optional, because a systematic review may include one but does not have to
- Narrative review
- does not necessarily follow a protocol, so how its studies were chosen is harder to check
- Study designs
- a systematic review is defined by its methods, not by one study design
Exam traps
- A meta-analysis is not needed for a review to be systematic: a systematic review may include one but does not have to, for example when the studies are too different to pool sensibly.
- Including randomised controlled trials only does not make a review systematic, because a review is systematic through how it is done rather than which designs it includes, and Cochrane reviews (a well-known example of systematic reviews) can include non-randomised studies.
- Expert authorship does not make a review systematic: a summary built from an expert's own choice of papers has no explicit, reproducible method, so another reviewer could not follow it and reach the same conclusions.
Read the full notes
Tested in 8 SimpleMed+ questions
Practice it in Pharmacology, with the answer and a full explanation after every question.
Practice these questions