Is my meta-analysis topic valid and feasible?

A topic is methodologically valid when the studies that exist can answer the question you are asking. It is feasible when you can find, assess and pool them within the time you have.

This is the test that a gap-hunting approach most often skips. A question can be a gap precisely because the studies needed to answer it do not exist.

Five questions to test validity

Are there enough studies? Count the primary studies that match your question. Two can be pooled technically, but fewer than about five give an unreliable estimate of heterogeneity, and funnel plots for publication bias are generally not recommended below ten studies.
Are the studies comparable? Check that populations, interventions, comparators and outcomes are close enough to pool. If they differ a lot, a narrative synthesis may be more honest than a pooled number.
Is the outcome reported in a form you can pool? You need an effect estimate with a measure of precision, or the numbers to calculate one. If most studies report the outcome in different or incomplete ways, the pool will be small.
Are the studies good enough? Risk of bias tools such as RoB 2 for randomized trials and ROBINS-I for non-randomized studies show how far a pooled result can be trusted. A pool of studies at high risk of bias gives a precise but unreliable estimate.
Is the question coherent? Combining very different interventions or populations can hide real effects. Decide in advance how broad the question is and write it into the protocol.

Red flags that mean stop or rethink

Only one or two small studies exist.
Every study defines the outcome differently, or reports it as a p-value only.
One very large trial dominates the evidence, so a pooled estimate would just repeat it.
Almost all studies are at high risk of bias.
The question is so broad that the pooled result could not be interpreted, or so narrow that nothing is left.

Let SynthGap run the first pass

Enter an area. SynthGap proposes candidate topics, checks each against existing systematic reviews, and lists the primary studies behind it. The free plan includes 3 discovery runs a month.

Start free

Feasibility

A full review involves two reviewers screening titles, abstracts and full text, extracting data in duplicate, assessing risk of bias, and reporting to PRISMA 2020 standards. The certainty of the conclusion is usually graded with GRADE. Estimate the number of records you will have to screen before you commit.

Testing the topic with a small, quick pool before committing is sensible. SynthGap's deep analysis does this: it captures the related studies and runs a small meta-analysis so you can see whether the evidence is thin before you invest months.

Questions

Is a topic with only a few studies a bad choice?

Not always. A review of a few small trials can still be useful when no estimate exists, and the honest result is a pooled estimate with wide uncertainty. But it should say clearly that the evidence is thin.

What is the difference between validity and feasibility?

Validity is whether the existing studies can answer the question. Feasibility is whether you can carry out the review in the time and with the skills you have.

Do I need a protocol before I start?

Yes. A protocol fixes the question, the eligibility criteria and the methods before you see the results, which protects against choosing methods that favour a result.

Keep reading

General methodological guidance, not medical advice. SynthGap's suggestions are AI-generated from the published literature, so verify them before registering a protocol.