Systematic Reviews and Meta-Analysis in Healthcare Simplified
From protocol to publication, without a statistics PhD.
The problem it solves
A committee asks for a systematic review, or a DNP project requires one, and the only models available are finished Cochrane reviews running to fifty pages. They show the destination and nothing of the route.
How it works
The whole sequence in order — question, protocol, search, screening, extraction, appraisal, synthesis, and the meta-analysis itself — at a level that assumes clinical training rather than research training.
Who it is for
- DNP and doctoral candidates whose project requires a systematic review
- Clinicians contributing to guideline development
- Anyone who has been asked for a review and shown a finished Cochrane one as the model
Who should look elsewhere. A narrative or scoping review answers a different question and needs a different method. If you are collecting primary data, this is the wrong book — though it is a reasonable place to start, since a proper review tells you whether the primary study is needed at all.
What is inside
- A focused, answerable question, and a protocol registered before you start
- A reproducible search strategy across the databases your field expects
- Screening, with the decisions documented well enough to survive scrutiny
- Data extraction that does not have to be redone
- Risk-of-bias appraisal you can defend
- Synthesis — and an honest judgement about whether the data support pooling
- Meta-analysis where they do, including heterogeneity and its interpretation
- GRADE, and stating certainty rather than implying it
- A PRISMA-compliant manuscript ready to submit
What you will have finished
- A focused, answerable review question and a registered protocol
- A reproducible search strategy and a documented screening process
- Extraction and risk-of-bias appraisal you can defend
- A synthesis, and a meta-analysis where the data support one
- A PRISMA-compliant manuscript ready to submit
What this adds to the free guidance
The usual model available to a first-time reviewer is a finished fifty-page Cochrane review, which shows the destination and nothing of the route. This walks the sequence in order, at a level that assumes clinical training rather than research training, and is explicit about the decision most reviews get wrong — whether the studies are similar enough to pool at all.
Built on PICO, PRISMA, GRADE. Original tools that operationalise the published standards and point to the free official sources.
Frequently asked questions
What is the difference between this and a scoping review?
A systematic review answers a focused question with a pre-specified method and, where appropriate, a pooled estimate. A scoping review maps what exists on a broader question and does not usually appraise or pool. Choosing the wrong one is expensive, so the first chapter makes you decide before you search.
How many databases do I need to search?
Enough that a reader could believe you did not miss the obvious, which in health research usually means at least the major bibliographic databases for your field plus a trials register, and often grey literature. What matters more than the count is that the strategy is documented well enough to be re-run.
What does GRADE actually do?
It separates the strength of a recommendation from the certainty of the evidence behind it, and forces you to state the second explicitly. Reviews that skip it tend to imply more confidence than their evidence supports, which is what a careful reader will challenge first.
How long does a systematic review take?
Longer than almost anyone plans for. Screening and extraction dominate, and both scale with the size of your search rather than with the number of studies you finally include. A tightly framed question is the single biggest time saver, which is why the question chapter comes first.
Do I have to register the protocol?
You should, and for many journals and programmes you must. Prospective registration on PROSPERO or an equivalent is what distinguishes a systematic review from a literature search written up afterwards, and it protects you against the accusation of having changed the question once you saw the results.
Do I need a meta-analysis?
Only if the studies are similar enough in population, intervention and outcome to make a pooled estimate meaningful. Pooling heterogeneous studies produces a precise number that means nothing. A systematic review with a narrative synthesis and a clear explanation of why pooling was inappropriate is the stronger paper.
Can I do one on my own?
Screening and extraction are normally done in duplicate, and most journals expect at least two reviewers for the screening stage. Plan for a second person even if the rest of the work is yours.