Most students choose the label before the method. They decide they are writing a systematic review, then work out what that involves somewhere around the second week of searching.
The paper pays for it later. A reviewer asks how the eligibility criteria were decided, and there is no good answer, because they were decided after the searching had already shaped them.
So this article covers:
- What a systematic review is
- What a narrative review is
- The differences that actually matter when you sit down to write
- Scoping, umbrella, rapid, integrative
- Which one to pick, which is the part nobody really teaches you
What is a Literature Review?
A literature review summarizes published work on a topic. No recruitment, no ethics approval, no data collection. The studies already exist, and your job is to make sense of them. This part is simple to understand.
Confusion starts when students assume there is only one recipe. Two researchers can review the same topic, follow completely different methods, and both be right, because they were answering different questions.
Say you want to know whether tranexamic acid reduces bleeding in postpartum hemorrhage. Narrow question, trials behind it, a systematic review fits. Now say you want to explain how the coagulation cascade fails in postpartum hemorrhage. Same clinical territory, different paper entirely, and a systematic review would be the wrong tool for it.
Systematic Review vs Narrative Review at a Glance
| Feature | Systematic Review | Narrative Review |
| Main purpose | Answer a focused research question | Provide an overview or interpretation of a topic |
| Search strategy | Structured and comprehensive | Often less standardized |
| Eligibility criteria | Predefined | May be flexible |
| Study selection | Systematic | Usually author driven |
| Risk of bias assessment | Usually expected | Often not formally required |
| Data synthesis | Structured, with meta-analysis when appropriate | Narrative interpretation |
| Reproducibility | Higher | Lower |
| Best for | Focused evidence questions | Broad conceptual understanding |
| Time required | Higher | Lower |
| Methodological rigor | High and prespecified | More flexible |
A warning about that table. Narrative reviews come out looking weaker in every single row, and that reading is wrong. They are not diluted systematic reviews. They do something a systematic review cannot do, which is explain a broad subject to someone who does not yet understand it.
PRISMA causes the other common mistake. PRISMA 2020 is a reporting guideline, 27 items plus a flow diagram, and it governs what you disclose about a review you have already completed. Students cite it as though it were the method they followed. It is not. Your methodology comes from Cochrane or JBI. PRISMA only shapes the write up.
What is a Systematic Review?
Definition
A systematic review answers one specific question using a method fixed in advance and written down before any searching begins. The test is simple: could another researcher follow your protocol and finish with the same included studies?
Key characteristics
Most of them involve:
- a question narrow enough to actually answer
- eligibility criteria locked in before screening
- a documented search across PubMed, Embase, the Cochrane Library, and often more
- two people screening independently, with a third resolving disagreements
- standardized extraction
- risk of bias assessment (RoB 2 for randomized trials, Newcastle-Ottawa for observational designs)
- synthesis, narrative or statistical
That list is the short version of a process that runs several months. How to write a systematic review and meta-analysis covers each stage in the order you will actually meet them.
When should you conduct a systematic review?
When the question is narrow and the primary studies already exist. Effectiveness questions qualify. So do diagnostic accuracy questions: how well does a D-dimer cutoff rule out pulmonary embolism? Risk factor questions work too.
What does not qualify is a question nobody has studied yet. Spending six weeks screening to report that two small trials exist is a painful way to learn that lesson.
JBI’s current guidance sets out separate methodologies for different forms of evidence synthesis, which should tell you there is no single process covering every topic.
What Is a Narrative Review?
Definition
A narrative review summarizes and interprets the literature without a prespecified selection process. You read widely, keep what explains the topic best, and build an overview from it.
Main characteristics
- broad topic, conceptual question
- the search is flexible and stops once you have enough
- interpretation carries the paper
- trials, guidelines, and textbook physiology can sit in the same paragraph
- none of the prespecified machinery listed above
For international medical graduates in particular this is often the sensible first project. Why a narrative review is a smart research choice for IMGs covers the timeline and what journals expect.
When is a narrative review useful?
A narrative review is a type of literature review that summarizes, interprets, and critiques existing research on a topic using the author’s expertise rather than a rigid, systematic protocol. Pulling a broad clinical concept together for readers outside your specialty.
Medical students ask whether any of this counts as a real paper. It does. Narrative reviews are peer reviewed, indexed, and widely cited. Just do not describe one as though it carried the reproducibility of a systematic review, because that claim will not survive peer review.
Systematic Review vs Narrative Review: Key Differences
Research question
Systematic: focused, answerable, usually framed with PICO. Narrative: broader. Something closer to “what do we currently understand about this condition?”
Framing a question tightly enough to answer is the skill everything else depends on. Our basic research methodology course covers PICO framing and the alternatives for diagnostic and qualitative questions.
Search strategy
A systematic review records everything. Databases, search terms, date limits, the lot. Anyone can rerun it. Narrative searching is looser and often is not reported in full.
Recording a search well enough that someone else can rerun it takes practice. Using PubMed effectively for medical research covers MeSH terms, Boolean operators and field tags.
Study selection
Criteria first, then screening. That is the systematic rule, and it is not negotiable. In a narrative review you keep far more discretion over what makes the cut.
Quality appraisal
Formal risk of bias assessment is standard in systematic reviews. A narrative review might comment on the strength of a particular study, but nobody expects structured appraisal of every paper cited.
Data synthesis
Structured synthesis, plus meta-analysis when studies are similar enough to pool. Narrative reviews synthesize through the writing itself.
Pooling brings its own decisions, starting with which model to use. Fixed effect versus random effects meta-analysis covers when studies are similar enough to combine and what happens when they are not.
Reproducibility
The clearest line between them. A systematic review is built so another team can repeat it. A narrative review reflects one author’s reading, which is far harder to replicate.
Time and resources
More planning, more screening hours, usually more than one reviewer. Narrative reviews move faster and a smaller team can manage them.
The takeaway: systematic review for a structured synthesis answering a focused evidence question, narrative review for broader interpretive understanding.
Narrative Review
A well executed narrative review beats a botched systematic one every time.
Flexible does not mean unstructured. Scoping the topic, organising by theme instead of by paper, saying where the evidence is weak, and stating your approach so reviewers can follow it. A realistic first publication that still holds up.
Other Types of Literature Reviews
Those two are not your only options.
Scoping review
Scoping reviews map what is out there. They identify concepts, show how research in an area has been conducted, and locate gaps. The question is “what evidence exists?” rather than “does this work?”
JBI’s scoping review methodology covers protocol development, search strategy, extraction, and how results are presented. Still a structured process, in other words, just aimed at breadth.
Umbrella review
An umbrella review synthesizes existing systematic reviews and meta-analyses instead of primary studies. Useful when a topic already carries five or six systematic reviews and nobody is sure what to make of them collectively.
Rapid review
A rapid review compresses the process to deliver evidence faster, usually because a guideline committee or policy team needs an answer in weeks rather than a year. Cochrane publishes dedicated guidance for these, precisely because shortening the process involves trade offs that should be stated openly rather than buried in the limitations paragraph.
Integrative review
Integrative reviews combine different forms of evidence, experimental and non experimental studies plus theoretical literature, into a broader picture. Common in nursing research.
How to Choose the Right Review Type
Start from your goal and work backward.
| Your research goal | Consider |
| Answer a focused clinical or intervention question | Systematic review |
| Map what research exists on a topic | Scoping review |
| Summarize existing systematic reviews | Umbrella review |
| Provide evidence quickly for a decision | Rapid review |
| Understand a broad topic or concept | Narrative review |
| Integrate multiple forms of evidence | Integrative review |
Before choosing your methodology, ask
- What exactly is my question?
- Focused, or exploratory?
- What evidence exists already?
- Do I need a comprehensive documented search?
- Is formal risk of bias assessment necessary here?
- Has someone published a recent review on this?
- How much time, how many co-reviewers, how much methodological support?
That last question matters more than beginners expect. One reviewer, no protocol, a two week deadline: that is not a systematic review, whatever the title says.
JBI publishes a decision tree for working out whether a scoping review is the right choice, covering available evidence, knowledge gaps, concept clarification, and whether a synthesis already exists. Use it. Safer than copying whatever your supervisor’s last paper did.
Evidence Synthesis: Where Do Different Review Types Fit?
Evidence synthesis covers any method that brings findings from multiple studies together. Every review type above sits inside it.
The temptation is to stack them into a hierarchy with systematic reviews on top. Resist it. The right review type depends on your question and objective, not on which method sits highest in an evidence pyramid.
Two examples make the point. A scoping review mapping an emerging field is more useful than a systematic review that finds three heterogeneous studies and cannot pool them. And a rapid review that reaches the committee before the vote beats a rigorous one published after the decision.
Think in terms of function. Each review type answers a different kind of question, and the skill lies in matching the method to the question you actually have.
Basic Methodology for a Systematic Review
1. Define the question
PICO works when you have a population, intervention, comparator, and outcome. Diagnostic and qualitative questions need other frameworks.
2. Write a protocol
Methods first, review second. Register on PROSPERO where eligible. Prespecification is what stops your decisions drifting toward the results you happen to like.
Registration is its own small process and the form asks for more than most people expect. Registering a systematic review on PROSPERO covers every field and the reasons records get sent back.
3. Build the search
Databases, keywords, controlled vocabulary such as MeSH, Boolean operators. Record the full string for each database you search.
4. Screen
Titles and abstracts, then full texts. Log exclusion reasons at the full text stage. Rayyan and Covidence make this manageable.
5. Extract data
One standardized form, applied the same way to every study.
6. Assess risk of bias
Choose a tool that matches the study designs you included.
7. Synthesize
Narrative synthesis always. Meta-analysis only when populations, interventions, and outcomes are close enough to combine meaningfully.
8. Report
PRISMA 2020, checklist and flow diagram.
Systematic Review
Eight steps on the page. Roughly nine months in practice, and step two is where most reviews die.
Prespecification is what stops your decisions drifting toward the results you happen to like, and almost nobody gets a protocol right unsupervised. Work through PICO framing, PROSPERO registration, dual screening and synthesis with a mentor who has taken a review to publication.
Basic Methodology for a Narrative Review
Flexible does not mean unstructured. Reviewers still want reasoning they can follow.
1. Set the scope
Decide what is in and what is out. “Iron deficiency anemia in pregnancy” is a scope you can write about. “Anemia” is a very broad term, you can write a book about it.
2. Search
At least one major database, usually PubMed, then follow the citations of the papers you find. Exhaustive is not required. Stopping at ten results is a problem.
3. Read and select
Recent primary studies, existing reviews, guidelines. Track where findings agree and where they clash, because the disagreement is usually your most useful material.
4. Organize
Group by theme, mechanism, or timeline. Not by paper. One study per paragraph reads like a bibliography with connectives.
5. Write the synthesis
Say what the evidence collectively suggests. Then say where it is weak.
6. State your approach
A short note on which databases you searched and what you prioritized. Two sentences will do. It improves transparency, and reviewers notice.
Common Mistakes Students and Young Researchers Make
Choosing a systematic review because it sounds prestigious: Prestige is not a methodology. A well executed narrative review beats a botched systematic one every time.
Calling a literature summary a systematic review: No protocol, no documented search, no eligibility criteria, no systematic review.
Treating PRISMA as the method: PRISMA is reporting. Conduct comes from Cochrane or JBI.
Searching before setting eligibility criteria: Search first and your criteria quietly reshape themselves around what you already found.
Picking the review type before defining the question: Always decide the question first. Methodology will follow.
Skipping the check for existing reviews: Look for a published synthesis or a registered protocol before you commit months to this.
Frequently Asked Questions
Q1. Is a systematic review better than a narrative review?
Neither wins outright. Systematic reviews are stronger for focused evidence questions. Narrative reviews suit broad concepts and emerging topics.
Q2. What is the difference between a systematic and narrative literature review?
A systematic review uses a predefined, reproducible method: documented searching, set criteria, formal appraisal. A narrative review searches flexibly and leads with interpretation.
Q3. Should a beginner conduct a systematic review?
Yes, with a supervisor and at least one co-reviewer. Screening and appraisal are difficult to do alone at an acceptable standard.
Q4. When should I choose a scoping review instead of a systematic review?
When you want to map evidence, clarify concepts, or find gaps rather than answer a narrow effectiveness question.
Q5. Is a narrative review considered research?
Yes. Peer reviewed publications that contribute interpretation and synthesis.
Q6. What is the difference between a scoping review and systematic review?
Scoping maps breadth. Systematic answers a specific question and usually includes risk of bias assessment.
Q7. What is an umbrella review?
A synthesis of existing systematic reviews and meta-analyses rather than primary studies.
Q8. What is the fastest type of literature review?
A rapid review, which streamlines parts of the systematic process to deliver evidence sooner.
Q9. Which review type is best for medical students?
Narrative or scoping, usually. Both fit a student timeline and still build real searching skills.
Conclusion
Pick the method that matches the question. Focused evidence question: systematic review, with a protocol, a documented search, and formal appraisal. Broad conceptual question: narrative review. Scoping, umbrella, rapid, and integrative reviews fill the space in between. Define the question first, and let the method follow from it, never the other way around.
Research Methodology
Define the question first, and let the method follow from it.
Every mistake listed above starts in the same place: a label chosen before a question was settled. Framing a question that can actually be answered, checking whether a synthesis already exists, and matching the design to the evidence in front of you are learnable skills, and they come before any review type.
The American Academy of Research & Academics offers mentorship and training for students and early career researchers, from your first research question through protocol development, search strategy, screening and manuscript preparation. If you are planning your first review, this is where it starts.
Disclaimer:
Articles published by American Academy of Research & Academics are prepared by our team using information from direct experience, publicly available resources, and educational references. AI tools may be used to assist with drafting, proofreading, and formatting; however, all content undergoes review and approval before publication.
The information provided is intended for educational purposes only. Requirements, policies, and processes may change over time. Readers should consult official sources for the most current information.