Skip to content

    Systematic review versus meta-analysis

    Systematic review versus meta-analysis

    One study says a treatment helped. Another finds a smaller effect. A third raises a question about side effects. How do you make sense of the whole collection without picking the headline you like best?

    A systematic review and a meta-analysis are two related tools for bringing research together. A systematic review uses a planned process to find and assess studies answering a specific question. A meta-analysis is a statistical way to combine compatible numerical results. A review may include one, but the terms don't mean the same thing.

    Begin with a question

    A useful review starts by defining what it wants to know. Researchers might ask how a particular treatment compares with another approach for a defined group of people, over a specified period, using a particular outcome.

    They then set criteria for which studies belong in the review. Those criteria may cover who participated, what was compared, the type of study and which outcomes were measured. A clear question helps readers understand why one study was included and another wasn't.

    The team searches for relevant evidence, selects studies according to its criteria, extracts information and assesses limitations. The methods should be described so readers can see how the review was done. “Systematic” describes the approach; it isn't a promise that every existing study was found or that all included studies were high quality.

    Where a meta-analysis fits

    If the studies are sufficiently comparable and the necessary data are available, reviewers may statistically combine results to produce a pooled estimate. That calculation is a meta-analysis. In a 2026 review of medicines for adults with overweight or obesity, researchers used a network meta-analysis drawing on 262 randomized trials. Network methods can use direct and indirect comparisons; the result still depends on which studies and participants were included, and on how the comparisons were made.

    Imagine three fictional studies comparing the same two approaches and measuring a similar outcome. A pooled result may summarize their estimates while taking account of how much information each contributes. It can provide a more precise summary than looking at one small study alone, but only if combining the studies makes sense for the question.

    The review should still show the individual study results and explain differences among them. A pooled number doesn't make the underlying studies identical or erase their limitations.

    Not every review should pool the numbers

    Studies may use different treatments, participants, outcomes or follow-up periods. One may measure a symptom after a few weeks while another examines a different score a year later. Important data may be missing, or the methods may differ enough that a single combined estimate would be hard to interpret.

    In those circumstances, a systematic review can still describe and compare the evidence without producing one pooled number. That isn't automatically a weaker review; it may be a more honest reflection of what the studies allow.

    The reverse also matters: a statistical combination isn't automatically a systematic review. Ask how the studies were found and selected, not just whether the report contains a forest plot or a pooled percentage.

    Quality in, quality out

    A review inherits questions from its underlying studies. If many trials have missing data, inconsistent methods or other risks of bias, combining their results doesn't make those problems disappear. The review team should assess those limitations and explain how they affect confidence in the conclusion.

    Missing studies matter too. If findings that are harder to access differ from those in published reports, the available picture may be incomplete. Reviewers may search trial records and other sources, but they can't always recover every result.

    Check the review date as well. A well-conducted review summarizes the evidence available when its search was completed; relevant studies published afterward may change the picture. Ask whether an update or newer evidence exists.

    Read the result in context

    Look for the estimated effect and its confidence interval, the number and size of studies, and the outcome being discussed. A review might find a modest average benefit for one outcome while the evidence for another remains uncertain.

    Consider whether the participants and comparison reflect your situation. A pooled result from people receiving a treatment plus a structured support program may not answer exactly the same question as the treatment alone. Follow-up length and side effects also belong in the discussion.

    If individual studies point in different directions, ask whether the review explored why. Differences in people, treatments or measurement may be relevant; sometimes the reason remains unclear. A single average can hide those differences.

    Questions to bring to an appointment

    You might ask: “What question did the review address? How many studies were included? Were they similar enough to combine? How certain are the results, and have newer studies appeared?” A clinician can help connect the review's population and outcomes to your own care.

    Don't treat the word “meta-analysis” as a guarantee or dismiss a review simply because it didn't pool results. Both the search and the judgment behind the synthesis matter.

    Think of a review as a map

    A systematic review tries to show the available evidence for a defined question and its limitations. A meta-analysis may add a numerical summary when combining results is appropriate. Read the map's boundaries, not just the number printed in the middle, and it becomes a more useful guide for the next conversation.