A systematic review is a structured method to find, select and appraise all the studies that answer a defined research question. A meta-analysis is a statistical technique that combines the numerical results of several of those studies into a single estimate.
The two are often mentioned together, and for good reason: most meta-analyses in health technology assessment (HTA) are carried out inside a systematic review. But they are not the same thing, and the difference matters when evidence is presented to decision-makers.
What is a systematic review?
A systematic review answers a focused question, usually framed as a PICO: the Population, the Intervention, the Comparators and the Outcomes. It follows a protocol written before the search starts, so that the choices cannot be adjusted to the results.
The main steps are:
- Protocol. The question, eligibility criteria, databases and methods are defined in advance.
- Search. Several bibliographic databases and other sources are searched with a reproducible strategy.
- Selection. Two reviewers screen the records independently against the eligibility criteria.
- Data extraction and risk of bias. Study characteristics and results are extracted and each study is assessed for the quality of its methods.
- Synthesis. The results are summarised, in a narrative way or, when possible, with a meta-analysis.
- Reporting. The review is reported transparently, usually following the PRISMA 2020 statement.
Our article on systematic literature reviews in HTA explains why this process is the foundation of most HTA submissions.
What is a meta-analysis?
A meta-analysis pools the results of two or more studies that measured the same outcome, giving more weight to the larger and more precise studies. The result is an overall estimate of the treatment effect, with a confidence interval, and a measure of how much the studies disagree with each other (heterogeneity).
There are two common models:
- Fixed-effect model, which assumes that all studies estimate the same underlying effect.
- Random-effects model, which allows the true effect to vary between studies and is often preferred when populations or settings differ.
When the aim is to compare several treatments that were not all tested against each other, a network meta-analysis combines direct and indirect evidence across a network of trials. Read more in our article on indirect treatment comparisons in HTA.
The key differences
- Nature. A systematic review is a research method; a meta-analysis is a statistical analysis.
- Result. A systematic review can end with a narrative summary; a meta-analysis always produces a pooled numerical estimate.
- Dependency. A systematic review does not need a meta-analysis. A meta-analysis should be based on a systematic review, otherwise the choice of studies may be biased.
- Feasibility. A meta-analysis is only appropriate when the studies are similar enough in population, intervention, comparators, outcomes and design to be combined.
In short: every good meta-analysis sits on a systematic review, but not every systematic review ends with a meta-analysis.
When is a meta-analysis not appropriate?
Pooling results is not always meaningful. A meta-analysis should be avoided, or interpreted with great caution, when:
- the studies define or measure the outcome in different ways;
- the populations or treatments are too different to answer the same question;
- heterogeneity is high and cannot be explained;
- most studies have a high risk of bias.
In these cases, a structured narrative synthesis is the honest choice. Assessing this feasibility before any statistics are run is part of good practice.
Why the distinction matters in HTA
HTA bodies expect the comparative evidence to come from a transparent and complete search. At European level, the methodological guidelines of the Member State Coordination Group on HTA assume that evidence synthesis starts from a properly conducted systematic literature review. A meta-analysis built on a selective set of studies will lose credibility in the assessment.
A systematic review also shows where the evidence is missing. That is the starting point of an evidence gap analysis, which helps companies plan the studies and analyses that an HTA submission will need.
Frequently asked questions
Can a meta-analysis be done without a systematic review?
Technically yes, but the result may be biased by which studies were chosen. In HTA, a meta-analysis is expected to be based on a systematic search and selection.
Is an umbrella review the same as a meta-analysis?
No. An umbrella review is a review of existing systematic reviews and meta-analyses. It is useful when many reviews already exist on a topic, as in our umbrella review on statin safety.
How long does a systematic review take?
It depends on the scope of the question and the volume of evidence. A focused review is much faster than one with several comparators and outcomes, so the PICO should be defined carefully from the start.
Final thoughts
A systematic review answers the question “what is all the evidence, and how good is it?”. A meta-analysis answers “what is the combined effect, and how certain are we?”. Used together and with transparent methods, they give HTA bodies and clinicians the reliable evidence they need.
At Clevidence, we conduct systematic, scoping and umbrella reviews, conventional and network meta-analyses and indirect treatment comparisons. See how we bridged an evidence gap with indirect comparisons, or learn more about our evidence review and synthesis services.