Everything in a systematic review traces back to the question. It decides your search terms, your eligibility criteria, your analysis, and whether the finished review answers anything worth knowing. Get it right and the project has a clear target. Get it wrong, too broad, too narrow, or unanswerable, and no amount of careful screening will rescue it. You cannot out-method a bad question.
This is separate from choosing a framework like PICO. The framework is the structure you pour the question into. This is about whether the question itself is any good in the first place.
A good systematic review question is focused, answerable, feasible, and meaningful. It targets a specific gap, it is narrow enough that you can decide which studies are relevant, there is enough evidence to actually address it, and answering it would be useful to someone. A framework such as PICO then gives that question a formal structure.
The failure modes are predictable: questions so broad they cannot be searched cleanly, questions so narrow no studies exist, and questions that were already answered by an existing review.
A useful checklist for research questions in general is FINER: Feasible, Interesting, Novel, Ethical, Relevant. It applies neatly to systematic reviews.
Feasible. There has to be enough evidence to review, and you have to have the time and team to do it. A question with three eligible studies rarely supports a systematic review; a question with fifteen thousand records may be more than a small team can screen. Feasibility sits between those extremes.
Interesting and relevant. The answer should matter to someone: clinicians, policymakers, other researchers. A technically sound review of a question no one cares about is wasted effort. Ask who will use the answer and what they will do with it.
Novel. Check that the question has not already been answered by a recent, good-quality systematic review. If one exists, either your question needs a different angle or you are looking at an update rather than a new review. This check happens early, through scoping searches.
Ethical. Less of a constraint for reviews than for primary research, but it still applies to how you handle and report the evidence.
Above all, a good question is focused. "What helps depression" is a topic. "In adults with major depressive disorder, does aerobic exercise reduce depressive symptoms compared with usual care" is a question. The difference is specificity, and specificity is what makes a question answerable.
Scope is where most questions go wrong, in one of two directions.
Too broad. A sprawling question drags tens of thousands of records into screening, mixes together studies that should not be compared, and produces a review that either takes years or says nothing precise. If your question spans many populations, several interventions, and a grab-bag of outcomes, it is too broad. Narrow it by tightening the population, specifying the intervention, or focusing on the outcomes that matter most.
Too narrow. Over-tighten and you find almost no eligible studies, which leaves nothing to synthesize. If a preliminary search turns up only a handful of papers, the question may be too specific, or the evidence may simply not exist yet. Widen the population, relax an overly strict criterion, or reconsider whether a systematic review is the right tool.
The target is a question narrow enough to be answerable and broad enough to have evidence behind it. You usually cannot judge this from the armchair. You find it by testing.
The single best way to calibrate a question is a preliminary search, sometimes called a scoping search, before you finalize anything. Run a quick, rough search on your draft question and look at what comes back.
If you get an unmanageable flood of results, the question is too broad. If you get almost nothing, it is too narrow or the evidence is missing. If a recent systematic review already answers it, you have your novelty check. This early search is not your formal, documented search; it is reconnaissance, and it routinely reshapes the question before any real work begins. Writing a question, testing it, and adjusting it two or three times is normal and healthy. Committing to an untested question is how projects go wrong.
Once the question is focused, answerable, and worth asking, give it formal structure with a framework. For intervention questions that is usually PICO: Population, Intervention, Comparison, Outcome, which we explain in the PICO framework explained. For exposure, qualitative, or scoping questions, one of the PICO variants fits better.
The framework and the question quality are two different jobs. The framework ensures you have specified every element. Judging whether the question is good, feasible, novel, and worth answering, comes first and cannot be delegated to an acronym.
Starting from a topic, not a question. A topic is where you begin, not where you stop. If you cannot state your question in one specific sentence, you are not ready to search.
Never testing the scope. Committing to a question without a preliminary search means discovering it was too broad or too narrow after you have invested weeks. Test first.
Skipping the novelty check. Building a full review only to find a recent, good one already answered your question is a painful and avoidable waste. Search for existing reviews early.
Chasing an interesting question with no evidence. A fascinating question with three eligible studies is not a systematic review. Feasibility is not optional.
Letting the question drift. Once you have set and registered the question, changing it midstream to fit the results undermines the review. Refine before you commit, then hold to it.
What makes a good systematic review question? It is focused, answerable, feasible, and meaningful. It targets a specific gap, is narrow enough to screen studies against, has enough evidence behind it, and produces an answer someone can use.
How do I know if my question is too broad? Run a preliminary search. If it returns an unmanageable number of records or spans many unrelated populations and outcomes, it is too broad. Tighten the population, intervention, or outcomes.
How do I know if my question is too narrow? If a preliminary search finds only a handful of studies, the question may be too specific or the evidence may not exist. Consider widening it or whether a systematic review is the right approach.
Should I use PICO to write my question? Use PICO, or a variant, to give your question structure once you know it is a good question. Judging whether the question is worth asking comes first.
How do I check my question has not already been answered? Search existing systematic reviews and registries early. If a recent, good-quality review already answers it, reframe the question or treat the project as an update.
A good systematic review question is focused enough to answer, feasible given the evidence, novel, and useful. The framework you choose gives it structure, but the framework cannot make a bad question good. Start from a specific question rather than a topic, test its scope with a preliminary search, check that no recent review already answers it, and only then lock it in.
The time you spend getting the question right pays back across every later stage. Once it is set, structure it with the PICO framework and move into your protocol.
Question settled and ready to build the review? Verflux takes you from search through screening and meta-analysis in one place.
All analytical features included in the free trial. No credit card, no installation, no R or Python.
Create free account