快速综述与系统综述的区别：何时选用及如何开展

What's new?

Now you can research with Unlimited Agent Usage

5 Nov 2026·Deepak Raghava·4 min read

Rapid Review vs Systematic Review: Differences, When to Use Each and How to Do One

How a rapid review differs from a systematic review, when each is the right choice, and how to run a rapid review step by step with AI support.

A systematic review can take a year or more. When a decision cannot wait that long, researchers use a rapid review: the same logic, with some steps deliberately simplified and reported openly.

In short: a rapid review follows systematic review methods but streamlines or skips some steps, such as searching fewer databases, limiting dates or languages, and using one reviewer with checks instead of two independent reviewers. It usually takes weeks to a few months rather than a year or more. Choose a rapid review when a decision is time-bound and you can state the shortcuts openly; choose a full systematic review when the answer must be as complete and unbiased as possible. Both can be run with the systematic review agent.

Rapid review vs systematic review at a glance

Aspect

Systematic review

Rapid review

---|---|---

Typical timeline

Many months to over a year

A few weeks to a few months

Question

Focused, often broad in scope

Narrow, tied to a specific decision

Several databases, grey literature, citation searching

Fewer, carefully chosen databases; limits on dates or languages

Screening

Two reviewers, independently, at each stage

One reviewer, with a second checking a sample or all exclusions

Data extraction

Two reviewers or one plus full verification

One reviewer, with verification

Risk of bias

Formal tool, usually in duplicate

Formal tool, often single reviewer with checks

Synthesis

Narrative or meta-analysis

Usually narrative

Reporting

PRISMA 2020

PRISMA 2020, stating every shortcut taken

How long a full review takes varies widely. An analysis of reviews registered in PROSPERO found an average of about 67 weeks from registration to publication (Borah et al., BMJ Open, 2017).

What a rapid review is (and is not)

Cochrane's Rapid Reviews Methods Group describes a rapid review as a form of evidence synthesis that streamlines or omits parts of the systematic review process to produce evidence quickly. It is still systematic: the question, search, criteria and decisions are planned and documented. It is not a quick literature review written from the first papers you find.

Cochrane has published practical guidance for rapid reviews (Garritty et al., 2021). Its common thread is to simplify where the risk of missing evidence is lowest and to report every simplification.

Key moments: 2:24 why a systematic review takes 12–18 months · 6:32 rapid review vs systematic review · 8:41 when to do a rapid review, and when not · 27:53 AI as a second reviewer for screening · 47:48 what if AI misses a key paper?

When to choose a rapid review

A rapid review is a good fit when:

A policy, clinical or funding decision has a fixed deadline.
You need to scope a topic before committing to a full systematic review.
A health emergency or new technology needs a timely summary of early evidence.
A course project or thesis chapter needs systematic methods in limited time.

Choose a full systematic review instead when:

The answer will set a clinical guideline or major policy.
The topic is contested and missing a few studies could change the conclusion.
You plan a meta-analysis that needs every eligible study.
Your journal or funder explicitly requires a full systematic review.

How to do a rapid review in 7 steps

Check whether a review already exists. Search for recent systematic and rapid reviews first. Updating or building on one may be faster than starting fresh.
Agree a narrow question. Use PICO (population, intervention, comparison, outcome) and confirm it with the people who will use the answer. See these PICO question examples.
Write a short protocol. Record your sources, limits and planned shortcuts before you start, so the shortcuts are decisions rather than accidents.
Search smartly, not everywhere. Pick the two or three databases that matter most for your field and justify any date or language limits.
Screen with checks. One reviewer can screen if a second reviewer checks a sample or all exclusions. AI can act as an extra screener, but a person must review every include and exclude decision.
Extract and appraise with verification. Use a predefined data extraction form and have a second person check key values.
Synthesise and report openly. Most rapid reviews use a narrative synthesis. Report with a PRISMA flow diagram and a limitations section listing every shortcut.

Where AI helps, and where it does not

AI can draft PICO questions and criteria, suggest search terms, screen titles and abstracts as a second reviewer, and pull data into extraction tables. That suits rapid reviews, where time is the constraint.

It does not remove the need for judgment. Check AI exclusions against the full text, verify every extracted number, and describe how AI was used in your methods, as the webinar covers at 47:12.

Watch: the full systematic review process in 8 steps

To see what a rapid review shortens, it helps to watch the full process. Dr. Faheem Ullah walks through his 8-step systematic review in this SciSpace webinar.

注意转载授权 · 阅读原文